Distributed machine room cold and heat source intelligent regulation and control method based on multi-modal perception

Abstract

The application provides a kind of distributed machine room cold and heat source intelligentization regulation and control method based on multi-modal perception, the method includes the following steps: obtaining cold and heat source side operation data, energy supply path pipe wall thermal response data and energy consumption terminal operation data, identifying cold and heat source output change event, determining the transmission time delay of energy supply path, cold and heat attenuation parameters and response form parameters, and establishing cold and heat migration data;According to the cold and heat migration data, the in-transit cold and heat contribution of historical control instruction is calculated, and the effective load to be compensated is determined in combination with the expected cold and heat load;Determine the cold and heat power distribution result and the source side action time sequence under the constraint of equipment operation, generate the control instruction of cold and heat source equipment, water pump and branch regulating valve, and issue at the corresponding time.The application can quantify the transmission lag, attenuation and diffusion process of cold and heat, reduce repeated adjustment and terminal overcooling or overheating, improve the matching degree of cold and heat arrival time and demand time, reduce system energy consumption and improve the stability of regulation and control.

Description

Technical Field

[0001] This invention relates to the field of cold and heat source control technology for computer rooms, and particularly to an intelligent control method for distributed computer room cold and heat sources based on multimodal sensing. Background Technology

[0002] As data centers, communication equipment rooms, and large computer rooms develop towards campus-based and distributed architecture, the application of centralized cooling or heating from multiple cold and heat source equipment rooms to different energy-consuming terminals via hydraulic transmission and distribution networks is gradually increasing. Such systems typically include chillers, heat pump units, boilers, cold and heat storage devices, pumps, and regulating valves. By adjusting the output of the cold and heat source equipment, the flow rate, and the opening of branch valves, each energy-consuming terminal receives the cooling or heating required for its operation. Because the equipment load, environmental conditions, and operating plans of different energy-consuming terminals vary, their cooling and heating demands exhibit significant temporal fluctuations and spatial differences. Therefore, the output distribution of cold and heat source equipment and the timing of equipment operation directly affect the stability of terminal temperatures and the system's energy consumption.

[0003] Existing methods for controlling cold and heat sources in computer rooms typically determine the output of cold and heat source equipment, pump frequency, and regulating valve opening based on the current temperature of the energy-consuming terminal, the temperature difference between supply and return water, or the predicted load. Some methods optimize by using fixed transmission delays or steady-state energy efficiency parameters. However, when cold and heat are transported from the cold and heat source to the energy-consuming terminal, they are affected by factors such as the length of the energy supply path, the internal volume of the pipe, the mass flow rate, the cold and heat losses along the way, and the thermal inertia of the pipe wall. This results in varying degrees of transmission lag, cold and heat attenuation, and response diffusion, affecting the control accuracy and energy utilization efficiency of the distributed cold and heat source system. Summary of the Invention

[0004] To address the shortcomings of existing technologies, the present invention aims to provide an intelligent control method for distributed data center cold and heat sources based on multimodal perception, which seeks to solve the technical problems of poor control accuracy and low energy utilization efficiency in existing technologies.

[0005] To achieve the above objectives, the present invention provides: an intelligent control method for distributed data center cold and heat sources based on multimodal sensing, applicable to a cold and heat source system including multiple cold and heat source data centers, multiple energy-consuming terminals, and a hydraulic transmission and distribution network connecting the cold and heat source data centers and the energy-consuming terminals, comprising the following steps: Acquire multimodal data, which includes source-side operation data of each cold and heat source room, pipe wall thermal response data of each energy supply path, and end-point operation data of each energy-consuming terminal. Based on the source-side operating data, identify cold and heat source output change events, and correlate the cold and heat source output change events with the pipe wall thermal response and heat exchange response in a time sequence to determine the cold and heat migration data of the corresponding energy supply path. Based on the terminal operation data, the expected cooling and heating load of each energy-consuming terminal in the future control time domain is determined, and based on the historical control instructions and the cooling and heating migration data, the in-transit cooling and heating contribution acting on the energy-consuming terminal in the future control time domain is determined, thereby obtaining the effective uncompensated load based on the expected cooling and heating load and the in-transit cooling and heating contribution. Based on the effective uncompensated load, determine the cooling and heating power allocated by each cooling and heating source room to each energy-consuming terminal and the source-side action sequence that matches the demand time of the energy-consuming terminal, generate corresponding control commands, and issue the control commands at the corresponding command issuance time.

[0006] According to one aspect of the above technical solution, the steps for identifying changes in the output of cold and heat sources specifically include: When the valve connection status of the target energy supply path remains unchanged, the mass flow fluctuation is less than the preset flow threshold, and the output change of the non-target cold and heat source is less than the preset power threshold, the start-up and shutdown of the target cold and heat source, the adjustment of the water supply temperature, or the adjustment of the output of the cold and heat source equipment are identified as identifiable cold and heat source output change events. If no cold or heat source output change event occurs within the preset identification period, a bounded disturbance is applied to the target cold or heat source's water supply temperature setpoint or the target energy supply path's mass flow rate, so that the water supply temperature and the terminal ambient temperature corresponding to the bounded disturbance are both within their respective allowable ranges, and the bounded disturbance is identified as a cold or heat source output change event.

[0007] According to one aspect of the above technical solution, the step of correlating the cold and heat source output change events with the pipe wall thermal response and heat exchange response in a time series to determine the cold and heat migration data of the corresponding energy supply path specifically includes: Extract the source-side enthalpy flow change sequence corresponding to the cold and heat source output change event, and time-align the source-side enthalpy flow change sequence with the pipe wall thermal response sequence and the terminal enthalpy flow response sequence of the corresponding energy supply path to determine the transmission delay, cold and heat attenuation degree and response diffusion degree of the energy supply path, so as to form the cold and heat migration data of the energy supply path.

[0008] According to one aspect of the above technical solution, the steps for determining the transmission delay of the power supply path specifically include: Calculate the source-side enthalpy change sequence based on the mass flow rate of the i-th cold and heat source and the supply and return water temperature difference: ; In the formula, Let be the change in enthalpy flow on the source side of the i-th hot and cold source at time t. The specific heat capacity of the transported medium. Let be the mass flow rate of the i-th heat source / cold source at time t. and These are the supply water temperature and the return water temperature, respectively. The average heating and cooling power output of the heat source before the change event; Calculate the equivalent thermal response sequence of the pipe wall based on the infrared temperature of the pipe section observed between the i-th heat source and the j-th energy consumption terminal: ; In the formula, To observe the equivalent thermal response of the pipe wall at time t, To observe the corrected pipe wall temperature of the pipe section, To observe the time constant of the pipe wall thermal response of the pipe section, This represents the rate of change of the pipe wall temperature relative to continuous time. For time differentiation; Within the time delay search interval determined based on the pipe volume and mass flow rate of the power supply path, the transmission delay is determined according to the following formula: ; In the formula, The transmission delay of the energy supply path between the i-th heat source and the j-th energy-consuming terminal; This represents the candidate time offset during the time-delay search process. and These are the lower and upper limits of the time delay search interval, respectively. and These represent the start and end times of the time series correlation analysis interval, respectively. and They are respectively and The first derivative relative to continuous time t, This represents the candidate time offset that maximizes the normalized correlation coefficient. The first derivative of the equivalent thermal response of the pipe wall is expressed as .

[0009] According to one aspect of the above technical solution, the steps for generating the cold and heat migration data of the energy supply path specifically include: The input heat or cold energy entering the target energy supply path is obtained by integrating the source-side enthalpy change sequence, the response heat or cold energy formed at the energy consumption end by the heat or cold source output change event is obtained by integrating the terminal enthalpy response sequence, and the response diffusion amount is obtained by the second-order central moment of the terminal enthalpy response sequence relative to the transmission delay. The cooling and heating attenuation parameters and response morphology parameters are determined based on the following formulas: ; ; In the formula, For the input of cold and heat to enter the corresponding energy supply path, The response heat or cold generated at the corresponding end, The second-order central moment of the terminal enthalpy flow response in the time dimension. This corresponds to the thermal attenuation parameters per unit transmission time along the power supply path. These are the response morphological parameters corresponding to the power supply path; The response kernel of the power supply path is established based on the transmission delay, thermal attenuation parameters, and response morphology parameters: ; In the formula, Let ω be the response kernel function value of the unit heating or cooling power on the source side when it reaches the j-th energy consumption terminal after time ω, where ω is the continuous elapsed time after the change in the output of the heating or cooling source.

[0010] According to one aspect of the above technical solution, the method further includes: When the valve connection status of the power supply path remains unchanged but the mass flow rate changes, the transmission delay and response morphology parameters are updated according to the following formula: ; ; In the formula, This represents the sequence number of the current discrete control cycle. The mass flow rate of the corresponding power supply path during the k-th control cycle. For reference quality flow rate, For reference transmission delay, For reference response morphology parameters, For the updated transmission latency, For the updated response morphology parameters; When the valve connection status changes, or when the deviation between the actual arrival time and the expected arrival time of the end exceeds the preset time delay threshold, the identification of cold and heat source output change events and the identification of transmission parameters are re-executed.

[0011] According to one aspect of the above technical solution, the steps for calculating the in-transit heating and cooling contribution and the effective uncompensated load specifically include: Based on the current enthalpy difference between supply and return water, changes in ambient temperature, and the terminal operation plan of the j-th energy consumption terminal, the expected cooling and heating load for the n-th control period is determined on a rolling basis. The system retrieves the cooling and heating power of the energy supply path from the i-th cooling and heating source to the j-th energy-consuming terminal within the historical control period, and calculates the in-transit cooling and heating contribution and the effective uncompensated load based on the following formula: ; ; In the formula, This represents the offset of the future control cycle relative to the current control cycle. The effective uncompensated load for the (k+n)th control cycle. Let j be the expected heating and cooling load of the j-th energy consumption terminal in the (k+n)-th control cycle. This represents the in-transit heating and cooling contribution generated by historical control commands during the (k+n)th control cycle. The total number of heat and cold sources. This is the offset from the historical control cycle relative to the current control cycle. To cover the historical control cycles within the effective time range of the corresponding response kernel, The duration of a single control cycle. For the i-th heat source in the i-th... Each control cycle input corresponds to the actual cooling and heating power of the energy supply path. To correspond to the historical hot and cold power after (n+ The response kernel function value applied to the j-th energy-consuming terminal after time Δ.

[0012] According to one aspect of the above technical solution, the steps for determining the distribution of cooling and heating power from each cooling and heating source room to each energy-consuming terminal and the source-side action sequence that matches the demand of the energy-consuming terminal specifically include: Using the cooling and heating power allocated by the i-th heat source to the j-th energy-consuming terminal in the future control cycle as the decision quantity, the expected cooling and heating power to be received at the terminal is calculated according to the following formula: ; In the formula, Let j be the expected cooling or heating power to be received by the j-th energy-consuming terminal in the (k+n)-th control cycle. This represents the offset of future control commands relative to the current control cycle. Let i be the heating or cooling power that the i-th heating or cooling source is intended to allocate to the j-th energy-consuming terminal in the k+ρ control cycle. The proposed allocation of heating and cooling power is processed through (n) The proportion of energy reaching the j-th energy-consuming terminal after time ρ)Δ; Under the conditions of satisfying the limits on the deviation of cold and hot power at the end of the energy consumption terminal, the upper and lower limits of the output of cold and hot source equipment, the rate of change of output of cold and hot source equipment, the shortest continuous operating time, the allowable flow rate of the water pump, the range of water supply temperature, and the mutual exclusion constraints of cold and hot conditions in the same hydraulic branch, the cold and hot power allocation result with the lowest input energy consumption shall be determined according to the following formula: ; In the formula, Let i be the measured input power function of the i-th heat source and cold source. Indicates that all time domains are controlled by the future. The set of decision variables to be optimized consists of J, which is the total number of energy-consuming terminals, and N, which is the number of control cycles contained in the future control time domain.

[0013] According to one aspect of the above technical solution, the step of generating corresponding control commands and issuing the control commands at the corresponding command issuance time specifically includes: The output of the cold and heat source equipment is determined by the sum of the cold and heat power distributed from the same cold and heat source to each energy-consuming terminal. The target mass flow rate is determined based on the output of the cold and heat source equipment, the temperature difference between the supply and return water, and the specific heat capacity of the transport medium. The target mass flow rate is converted into the pump operating frequency based on the pump performance curve. The opening degree of the branch regulating valve is determined based on the flow characteristics of the regulating valve and the pressure difference of the power supply path. The command issuance time for cold and heat source equipment, water pumps, and branch regulating valves is determined by subtracting the transmission delay of the corresponding energy supply path from the time of cold and heat demand at the end of the energy consumption terminal.

[0014] Compared with existing technologies, the beneficial effects of this invention are as follows: By identifying cold and heat source output change events, and combining source-side operating data, pipe wall thermal response data of the energy supply path, and energy-consuming terminal operating data, the transmission delay, cold and heat attenuation parameters, and response morphology parameters of the energy supply path between different cold and heat sources and different energy-consuming terminals are determined. Based on this, cold and heat migration data capable of characterizing the dynamic transmission characteristics of the path is established, enabling the quantification of the arrival time, attenuation degree, and time diffusion process of cold and heat in the transmission and distribution network. During the control process, the cold and heat migration data is used to calculate the in-transit cold and heat contribution generated by historical control commands in the future control period, and the in-transit cold and heat contribution is subtracted from the expected cold and heat load at the energy-consuming terminal. The system obtains the effective uncompensated load that still needs to be compensated, thereby reducing repeated adjustments caused by transmission lag. Furthermore, based on the effective uncompensated load and the transmission characteristics of different power supply paths, the system jointly determines the power distribution results of each cold and heat source and the timing of source-side actions under equipment operation constraints. This information is then converted into the output of cold and heat source equipment, the operating frequency of water pumps, and the opening of branch regulating valves. Control commands are issued at times that match the end-user demand, thereby improving the matching degree between the actual arrival time of cold and heat and the end-user demand, reducing end-user overcooling, overheating, and power fluctuations, reducing ineffective adjustments and frequent start-ups and shutdowns of cold and heat source equipment, and improving the control stability and energy utilization efficiency of the distributed data center cold and heat source system. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating an intelligent control method for distributed data center cold and heat sources based on multimodal sensing in one embodiment of the present invention. The following detailed description, in conjunction with the accompanying drawings, will further illustrate the present invention. Detailed Implementation

[0016] To facilitate understanding of the present invention, a more complete description will be given below with reference to the accompanying drawings. Several embodiments of the invention are illustrated in the drawings. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.

[0017] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.

[0018] Example 1 The first embodiment of this invention provides an intelligent control method for distributed data center cold and heat sources based on multimodal sensing. This method can be applied to a cold and heat source system including multiple cold and heat source data centers, multiple energy-consuming terminals, and a hydraulic transmission and distribution network connecting the cold and heat source data centers and the energy-consuming terminals. The cold and heat source data centers are equipped with one or more of the following: chillers, heat pumps, boilers, cold storage devices, and heat storage devices. The hydraulic transmission and distribution network is equipped with water pumps, regulating valves, supply water pipelines, and return water pipelines. A single cold and heat source data center can deliver cooling or heating energy to different energy-consuming terminals through one or more energy supply paths. The different energy supply paths are determined by the connection relationships of pipe sections and the connection status of valves along the path.

[0019] In this embodiment, an observation pipe section is installed on the side of the energy supply path closest to the energy consumption terminal, and the pipe wall temperature field of the observation pipe section is acquired using an infrared thermal imaging device. The observation pipe section is positioned before the water flow enters the energy consumption terminal, at a location that reflects the path arrival response and is not easily affected by the thermal radiation interference of the terminal heat exchange equipment itself. Temperature sensors, flow sensors, and pressure sensors can also be installed at the cold and heat source outlet, the energy consumption terminal inlet, and the energy consumption terminal outlet.

[0020] Please see Figure 1 The figure shows a flowchart of the intelligent control method for distributed data center cold and heat sources based on multimodal perception in the first embodiment of the present invention. As shown in the figure, the method includes the following steps: Step S100: Acquire multimodal data, which includes source-side operating data of each cold and heat source room, pipe wall thermal response data of each energy supply path, and terminal operating data of each energy-consuming terminal. In this embodiment, the acquired multimodal data includes: start-up and shutdown status of chillers, heat pumps, boilers, cold storage devices, or heat storage devices; output setpoints and input power of cold and heat source equipment; supply and return water temperatures, mass flow rates, pressure differentials, pump operating frequencies, and regulating valve openings on the source and terminal sides; infrared temperature field of the observed pipe section; and ambient temperature and operating plan of the energy-consuming terminal. The infrared temperature field is corrected according to the material type, coating state, and surface emissivity of the observed pipe section surface, and the pipe wall temperature sequence characterizing the overall temperature change of the observed pipe section is extracted from the corrected infrared temperature field.

[0021] Data from different acquisition devices are time-stamped according to a unified time base. The initial time stamp is the moment the operating command is issued, and the actual response time stamp is the moment when the input power of the cold and heat source devices changes in accordance with the operating command. This corrects for acquisition and communication delays in the water-side data and infrared temperature field sequences. After time delay correction, each data sequence is resampled according to a unified sampling period to ensure that source-side operating data, pipe wall thermal response data, and terminal operating data can be processed on the same time axis.

[0022] Step S200: Identify cold and heat source output change events based on the source-side operating data, and correlate the cold and heat source output change events with the pipe wall thermal response and heat exchange response in a time sequence to determine the cold and heat migration data of the corresponding energy supply path.

[0023] Specifically, after data acquisition and time correction, cold and heat source output change events are identified based on source-side operational data. When the valve connection status of the target energy supply path remains unchanged, the mass flow fluctuation is less than a preset flow threshold, and the output change of non-target cold and heat sources is less than a preset power threshold, the start-up and shutdown of the target cold and heat source, water supply temperature adjustment, or cold and heat source equipment output adjustment are identified as identifiable cold and heat source output change events. This reduces the interference of other cold and heat source output changes, flow abrupt changes, and valve switching on the path response identification results.

[0024] If no natural operating event meeting the above conditions occurs within the preset identification period, a bounded disturbance can be applied to the setpoint of the target cold / heat source's water supply temperature or the mass flow rate of the target energy supply path. The amplitude and duration of the bounded disturbance are limited to the equipment's allowable range, and the water supply temperature and the terminal ambient temperature during the disturbance process must not exceed their respective allowable ranges. The change in source-side cold / heat power corresponding to this bounded disturbance is taken as a cold / heat source output change event.

[0025] For the identified cold and heat source output change events, calculate the source-side enthalpy flow change sequence based on the mass flow rate of the i-th cold and heat source and the supply and return water temperature difference: ; In the formula, Let be the change in enthalpy flow on the source side of the i-th hot and cold source at time t. The specific heat capacity of the transported medium. Let be the mass flow rate of the i-th heat source / cold source at time t. and These are the supply water temperature and the return water temperature, respectively. The average heating and cooling power output of the heat source before the change event; For events where the output of cold and heat source equipment increases, the increased cold and heat power is taken as the positive enthalpy change; for events where the output of cold and heat source equipment decreases, the source-side enthalpy change sequence and the corresponding terminal enthalpy response sequence are converted in the same direction so that they have a consistent direction of change in correlation analysis and cold and heat integration.

[0026] Based on the corrected pipe wall temperature observed in the pipe section between the i-th heat source and the j-th energy consumption terminal, calculate the equivalent thermal response sequence of the pipe wall: ; In the formula, To observe the equivalent thermal response of the pipe wall at time t, To observe the corrected pipe wall temperature of the pipe section, To observe the time constant of the pipe wall thermal response of the pipe section, This represents the rate of change of the pipe wall temperature relative to continuous time. The time constant of the pipe wall thermal response can be calibrated based on the historical step response of the observed pipe section, and is used to compensate for the response lag caused by the thermal inertia of the pipe wall.

[0027] The nominal delivery time is determined based on the pipe volume of the power supply path and the current mass flow rate. Preset extension ranges are set on both sides of the nominal delivery time to form a time delay search interval. Within this time delay search interval, the transmission delay is determined according to the following formula: ; In the formula, The transmission delay of the energy supply path between the i-th heat source and the j-th energy-consuming terminal; This represents the candidate time offset during the time-delay search process. and These are the lower and upper limits of the time delay search interval, respectively. and These represent the start and end times of the time series correlation analysis interval, respectively. and They are respectively and The first derivative relative to continuous time t, This represents the candidate time offset that maximizes the normalized correlation coefficient. The first derivative of the equivalent thermal response of the pipe wall is expressed as .

[0028] Using the rate of change instead of directly using enthalpy flow and pipe wall temperature values ​​for correlation analysis can reduce the impact of differences in source-side reference power, ambient temperature, and initial pipe wall temperature on the correlation results. Normalization can reduce the scale differences caused by the output amplitude of different heat and cold source devices and the temperature rise amplitude of different observed pipe sections.

[0029] After determining the transmission delay, the input heat or cold energy entering the corresponding energy supply path is obtained by integrating the source-side enthalpy change sequence over the duration of the cold / heat source output change event. The terminal enthalpy response sequence is determined based on the terminal mass flow rate and the enthalpy difference between the supply and return water. This terminal enthalpy response sequence is then integrated over the response period to obtain the terminal response heating / cooling heat. .

[0030] Using the time of occurrence of the cold / heat source output change event plus the transmission delay as the expected arrival time of the center, the second-order central moment of the terminal enthalpy flow response sequence relative to this expected arrival time of the center is calculated to obtain the response diffusion. The greater the response diffusion, the higher the degree of dispersion of heat and cold over time after passing through bends, branches, and local mixing areas.

[0031] The cooling and heating attenuation parameters and response morphology parameters are determined based on the following formulas: ; ; In the formula, For the input of cold and heat to enter the corresponding energy supply path, The response heat or cold generated at the corresponding end, The second-order central moment of the terminal enthalpy flow response in the time dimension. This corresponds to the thermal attenuation parameters per unit transmission time along the power supply path. These are the response morphological parameters corresponding to the power supply path; The response kernel of the power supply path is established based on the transmission delay, thermal attenuation parameters, and response morphology parameters: ; In the formula, Let be the response kernel function value when the unit heat or cold power from the source side reaches the j-th energy consumption terminal after time ω, where ω is the continuous elapsed time after the heat or cold source output change event occurs. The response kernel is an empirical response function that characterizes the transmission delay, heat or cold attenuation, and response diffusion characteristics of the energy supply path. Its integral over the entire response time range represents the equivalent proportion of heat or cold power from the source side reaching the energy consumption terminal after transmission through the corresponding energy supply path.

[0032] Transmission delay, thermal attenuation parameters, response morphology parameters, and response kernels together constitute the thermal migration data for the corresponding energy supply path. Thermal migration data describes the dynamic response of the energy supply path between a defined heat source and a defined energy consumer. Since the supply and return water temperatures, heat dissipation direction, and equipment operating boundaries differ under cooling and heating conditions for the same energy supply path, cooling and heating events are collected separately, and separate response kernels are established for cooling and heating conditions. During control operations, the thermal migration data for the corresponding condition is retrieved based on the current energy supply mode, avoiding the direct use of cooling parameters for heating control or vice versa.

[0033] When the valve connection status of the power supply path remains unchanged but the mass flow rate changes, update the transmission delay and response morphology parameters based on the current mass flow rate: ; ; In the formula, This represents the sequence number of the current discrete control cycle. The mass flow rate of the corresponding power supply path during the k-th control cycle. For reference quality flow rate, For reference transmission delay, For reference response morphology parameters, For the updated transmission latency, The updated response morphology parameters, with the superscript 0 indicating the reference operating condition when identifying hot and cold migration data; When the valve connection status of the energy supply path changes, the pipe volume, branch connection relationship, or mixing position of the original energy supply path may change. Instead of directly updating the parameters using the mass flow rate ratio, the system re-identifies the cold and heat source output change events and re-executes the transmission parameter identification. When the deviation between the actual arrival time and the expected arrival time of the terminal exceeds the preset time delay threshold, the path identification is also re-performed.

[0034] Step S300: Determine the expected cooling and heating load of each energy-consuming terminal in the future control time domain based on the terminal operation data, and determine the in-transit cooling and heating contribution amount acting on the energy-consuming terminal in the future control time domain based on the historical control instructions and the in-transit cooling and heating contribution amount, thereby obtaining the effective compensation load based on the expected cooling and heating load and the in-transit cooling and heating contribution amount.

[0035] Specifically, during the regulation phase, the current control cycle is designated as the k-th control cycle, and the next N consecutive control cycles are defined as the future control time domain, with each control cycle lasting Δ. Based on the current enthalpy difference between supply and return water, changes in ambient temperature, and the terminal operation plan at the j-th energy consumption terminal, the expected cooling and heating load for the k+n-th control cycle is determined on a rolling basis. , where 1≤n≤N, and N is the number of control cycles contained in the future control time domain.

[0036] The actual cooling and heating power of the energy supply path from the i-th heat source to the j-th energy-consuming terminal within the historical control period is used, and the in-transit cooling and heating contribution is calculated according to the following formula: ; In the formula, This represents the offset of the future control cycle relative to the current control cycle. This represents the in-transit heating and cooling contribution generated by historical control commands during the (k+n)th control cycle. The total number of heat and cold sources. This is the offset from the historical control cycle relative to the current control cycle. To cover the historical control cycles within the effective time range of the corresponding response kernel, The duration of a single control cycle. For the i-th heat source in the i-th... Each control cycle input corresponds to the actual cooling and heating power of the energy supply path. To correspond to the historical hot and cold power after (n+ The response kernel function value applied to the j-th energy-consuming terminal after time Δ.

[0037] The heat and cold contributions en route correspond to the heat and cold that have entered the hydraulic transmission and distribution network but have not yet been fully reflected in the measured heat exchange at the end of the pipeline, and will continue to affect the energy consumption end in future control cycles. Subtracting the heat and cold contributions en route from the expected heat and cold load yields the effective uncompensated load: ; In the formula, Let j be the expected heating and cooling load of the j-th energy consumption terminal in the (k+n)-th control cycle. The effective uncompensated load for the (k+n)th control cycle is... This represents the effective uncompensated load for the k+n control cycle. By deducting the in-transit heating and cooling contributions, we can avoid repeatedly increasing the output of heating and cooling source equipment based solely on the current temperature deviation at the terminal, which could lead to overcooling or overheating at the terminal or frequent start-ups and shutdowns on the source side.

[0038] Step S400: Based on the effective uncompensated load, determine the cooling and heating power allocated by each cooling and heating source room to each energy-consuming terminal and the source-side action sequence that matches the demand time of the energy-consuming terminal, generate corresponding control commands, and issue the control commands at the corresponding command issuance time.

[0039] In this embodiment, the cooling and heating power to be allocated from the i-th heat source to the j-th energy-consuming terminal in the k+ρ-th control cycle is... For decision-making purposes, calculate the expected cooling and heating power to be reached at the terminal: ; In the formula, Let j be the expected cooling or heating power to be received by the j-th energy-consuming terminal in the (k+n)-th control cycle. This represents the offset of future control commands relative to the current control cycle. Let i be the heating or cooling power that the i-th heating or cooling source is intended to allocate to the j-th energy-consuming terminal in the k+ρ control cycle. The proposed allocation of heating and cooling power is processed through (n) The proportion density reaching the j-th energy-consuming end after time ρ)Δ.

[0040] Under the conditions of satisfying the limits on the deviation of cold and hot power at the end of the energy consumption terminal, the upper and lower limits of the output of cold and hot source equipment, the rate of change of output of cold and hot source equipment, the shortest continuous operating time, the allowable flow rate of the water pump, the range of water supply temperature, and the mutual exclusion constraints of cold and hot conditions in the same hydraulic branch, the cold and hot power allocation result with the lowest input energy consumption shall be determined according to the following formula: ; In the formula, Let i be the measured input power function of the i-th heat source and cold source. Indicates that all time domains are controlled by the future. The set of decision variables to be optimized consists of J, where J is the total number of energy-consuming terminals. The measured input power function can be established based on the historical input power and output cooling / heating power of the cold and heat source equipment under different output conditions, so as to reflect the actual operating efficiency of the equipment under different load rates.

[0041] After obtaining the power distribution results, the power distributed from the same heat source to each energy-consuming terminal is added together to determine the target output of the corresponding heat source equipment. The target mass flow rate is determined based on the target output of the heat source equipment, the supply and return water temperature difference, and the specific heat capacity of the transport medium. Then, the target mass flow rate is converted into the pump operating frequency according to the pump performance curve. Finally, the target opening degree of each branch regulating valve is determined based on the flow characteristics of the regulating valve and the pressure difference along the energy supply path.

[0042] For the energy supply path from the i-th heat source to the j-th energy consumer, the transmission delay of the energy supply path is subtracted from the time of the corresponding energy consumer's heating or cooling demand to obtain the command issuance time for the heat source equipment, water pump, and branch regulating valve. The controller generates corresponding control commands and issues control commands to the heat source equipment, water pump, and branch regulating valve at the corresponding command issuance time, so that heating or cooling reaches the corresponding end when demand is generated at the end.

[0043] Preferably, in some application scenarios of this embodiment, after the control command is executed, the actual enthalpy flow response at the terminal is obtained based on the actual mass flow rate at the terminal and the actual enthalpy difference between the supply and return water. The actual enthalpy flow response at the terminal is then compared with the expected enthalpy flow response determined based on the heat migration data and the issued control command. The arrival time deviation between the actual arrival time and the expected arrival time, the response heat deviation between the actual response heat and the expected response heat, and the response diffusion deviation between the actual response diffusion amount and the expected response diffusion amount are calculated respectively.

[0044] Calculate the correlation peak between the cold / heat source output change event and the actual enthalpy flow response at the terminal. When the correlation peak exceeds a preset confidence threshold, the current response is determined as a reliable response. The transmission delay is corrected based on the arrival time deviation, the cold / heat attenuation parameters are re-determined based on the actual input and actual response cold / heat, and the response morphology parameters are re-determined based on the actual response diffusion. When the arrival time deviation exceeds a preset delay threshold or the valve connection status changes, a complete cold / heat migration data identification process is re-executed.

[0045] When the relevant peak value is not greater than the preset confidence threshold, it indicates that the terminal response may be affected by changes in other cold and heat sources, sudden changes in terminal load, or sensor anomalies. At this time, the transmission parameters before the update are retained, the cold and heat migration data are not modified using this response, and conservative control is performed according to the cold and heat source equipment's allowed output, water supply temperature boundary, and terminal temperature boundary until a cold and heat source output change event and terminal response that meet the confidence conditions are obtained.

[0046] In summary, the intelligent control method for distributed data center cold and heat sources based on multimodal perception in the above embodiments of the present invention identifies cold and heat source output change events, combines source-side operating data, pipe wall thermal response data of the energy supply path, and energy-consuming terminal operating data to determine the transmission delay, cold and heat attenuation parameters, and response morphology parameters of the energy supply path between different cold and heat sources and different energy-consuming terminals. Based on this, cold and heat migration data that characterizes the dynamic transmission characteristics of the path is established, enabling the quantification of the arrival time, attenuation degree, and time diffusion process of cold and heat in the transmission and distribution network. During the control process, the cold and heat migration data is used to calculate the in-transit cold and heat contribution generated by historical control commands in the future control period, and the contribution is deducted from the expected cold and heat load of the energy-consuming terminals. The system calculates the on-transit heating and cooling contribution to obtain the effective load that still needs to be compensated, thereby reducing redundant adjustments caused by transmission lag. Furthermore, based on the effective load that needs to be compensated and the transmission characteristics of different power supply paths, the system jointly determines the heating and cooling power distribution results and source-side action sequence of each heating and cooling source under equipment operation constraints. This is then converted into the output of heating and cooling source equipment, the operating frequency of water pumps, and the opening degree of branch regulating valves. Control commands are issued at the time that matches the end-user demand time. This improves the matching degree between the actual arrival time of heating and cooling and the end-user demand time, reduces end-user overcooling, overheating, and heating and cooling power fluctuations, reduces ineffective adjustments and frequent start-ups and shutdowns of heating and cooling source equipment, and improves the control stability and energy utilization efficiency of the distributed data center heating and cooling source system.

[0047] The embodiments described above merely illustrate specific implementation methods of this application, and while the descriptions are detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A distributed data center cold and heat source intelligent control method based on multimodal sensing, applied to a cold and heat source system including multiple cold and heat source data centers, multiple energy-consuming terminals, and a hydraulic transmission and distribution network connecting the cold and heat source data centers and the energy-consuming terminals, characterized in that, Includes the following steps: Acquire multimodal data, which includes source-side operation data of each cold and heat source room, pipe wall thermal response data of each energy supply path, and end-point operation data of each energy-consuming terminal. Based on the source-side operating data, identify cold and heat source output change events, and correlate the cold and heat source output change events with the pipe wall thermal response and heat exchange response in a time sequence to determine the cold and heat migration data of the corresponding energy supply path. Based on the terminal operation data, the expected cooling and heating load of each energy-consuming terminal in the future control time domain is determined, and based on the historical control instructions and the cooling and heating migration data, the in-transit cooling and heating contribution acting on the energy-consuming terminal in the future control time domain is determined, thereby obtaining the effective uncompensated load based on the expected cooling and heating load and the in-transit cooling and heating contribution. Based on the effective uncompensated load, determine the cooling and heating power allocated by each cooling and heating source room to each energy-consuming terminal and the source-side action sequence that matches the demand time of the energy-consuming terminal, generate corresponding control commands, and issue the control commands at the corresponding command issuance time.

2. The intelligent control method for distributed data center cold and heat sources based on multimodal perception according to claim 1, characterized in that, The specific steps for identifying changes in the output of hot and cold sources include: When the valve connection status of the target energy supply path remains unchanged, the mass flow fluctuation is less than the preset flow threshold, and the output change of the non-target cold and heat source is less than the preset power threshold, the start-up and shutdown of the target cold and heat source, the adjustment of the water supply temperature, or the adjustment of the output of the cold and heat source equipment are identified as identifiable cold and heat source output change events. If no cold or heat source output change event occurs within the preset identification period, a bounded disturbance is applied to the target cold or heat source's water supply temperature setpoint or the target energy supply path's mass flow rate, so that the water supply temperature and the terminal ambient temperature corresponding to the bounded disturbance are both within their respective allowable ranges, and the bounded disturbance is identified as a cold or heat source output change event.

3. The intelligent control method for distributed data center cold and heat sources based on multimodal sensing according to claim 1, characterized in that, The steps of correlating the cold and heat source output change events with the pipe wall thermal response and heat exchange response in a time series to determine the cold and heat migration data of the corresponding energy supply path specifically include: Extract the source-side enthalpy flow change sequence corresponding to the cold and heat source output change event, and time-align the source-side enthalpy flow change sequence with the pipe wall thermal response sequence and the terminal enthalpy flow response sequence of the corresponding energy supply path to determine the transmission delay, cold and heat attenuation degree and response diffusion degree of the energy supply path, so as to form the cold and heat migration data of the energy supply path.

4. The intelligent control method for distributed data center cold and heat sources based on multimodal perception according to claim 3, characterized in that, The specific steps for determining the transmission delay of the power supply path include: Calculate the source-side enthalpy change sequence based on the mass flow rate of the i-th cold and heat source and the supply and return water temperature difference: ； In the formula, Let be the change in enthalpy flow on the source side of the i-th hot and cold source at time t. The specific heat capacity of the transported medium. Let be the mass flow rate of the i-th heat source / cold source at time t. and These are the supply water temperature and the return water temperature, respectively. The average heating and cooling power output of the heat source before the change event; Calculate the equivalent thermal response sequence of the pipe wall based on the infrared temperature of the pipe section observed between the i-th heat source and the j-th energy consumption terminal: ； In the formula, To observe the equivalent thermal response of the pipe wall at time t, To observe the corrected pipe wall temperature of the pipe section, To observe the time constant of the pipe wall thermal response of the pipe section, This represents the rate of change of the pipe wall temperature relative to continuous time. For time differentiation; Within the time delay search interval determined based on the pipe volume and mass flow rate of the power supply path, the transmission delay is determined according to the following formula: ； In the formula, The transmission delay of the energy supply path between the i-th heat source and the j-th energy-consuming terminal; This represents the candidate time offset during the time-delay search process. and These are the lower and upper limits of the time delay search interval, respectively. and These represent the start and end times of the time series correlation analysis interval, respectively. and They are respectively and The first derivative with respect to continuous time t, This represents the candidate time offset that maximizes the normalized correlation coefficient. The first derivative of the equivalent thermal response of the pipe wall is expressed as .

5. The intelligent control method for distributed data center cold and heat sources based on multimodal perception according to claim 4, characterized in that, The steps for generating the cold and heat migration data of the energy supply path specifically include: The input heat or cold energy entering the target energy supply path is obtained by integrating the source-side enthalpy change sequence, the response heat or cold energy formed at the energy consumption end by the heat or cold source output change event is obtained by integrating the terminal enthalpy response sequence, and the response diffusion amount is obtained by the second-order central moment of the terminal enthalpy response sequence relative to the transmission delay. The cooling and heating attenuation parameters and response morphology parameters are determined based on the following formulas: ； ； In the formula, For the input of cold and heat to enter the corresponding energy supply path, The response heat or cold generated at the corresponding end, The second-order central moment of the terminal enthalpy flow response in the time dimension. This corresponds to the thermal attenuation parameters per unit transmission time along the power supply path. These are the response morphological parameters corresponding to the power supply path; The response kernel of the power supply path is established based on the transmission delay, thermal attenuation parameters, and response morphology parameters: ； In the formula, Let ω be the response kernel function value of the unit heating or cooling power on the source side when it reaches the j-th energy consumption terminal after time ω, where ω is the continuous elapsed time after the change in the output of the heating or cooling source.

6. The intelligent control method for distributed data center cold and heat sources based on multimodal perception according to claim 5, characterized in that, The method further includes: When the valve connection status of the power supply path remains unchanged but the mass flow rate changes, the transmission delay and response morphology parameters are updated according to the following formula: ； ； In the formula, This represents the sequence number of the current discrete control cycle. The mass flow rate of the corresponding power supply path during the k-th control cycle. For reference quality flow rate, For reference transmission delay, For reference response morphology parameters, For the updated transmission latency, For the updated response morphology parameters; When the valve connection status changes, or when the deviation between the actual arrival time and the expected arrival time of the end exceeds the preset time delay threshold, the identification of cold and heat source output change events and the identification of transmission parameters are re-executed.

7. The intelligent control method for distributed data center cold and heat sources based on multimodal sensing according to claim 5, characterized in that, The specific steps for calculating the in-transit heating and cooling contributions and the effective uncompensated load include: Based on the current enthalpy difference between supply and return water, changes in ambient temperature, and the terminal operation plan of the j-th energy consumption terminal, the expected cooling and heating load for the n-th control period is determined on a rolling basis. The system retrieves the cooling and heating power of the energy supply path from the i-th cooling and heating source to the j-th energy-consuming terminal within the historical control period, and calculates the in-transit cooling and heating contribution and the effective uncompensated load based on the following formula: ； ； In the formula, This represents the offset of the future control cycle relative to the current control cycle. For the effective uncompensated load in the (k+n)th control cycle, Let j be the expected heating and cooling load of the j-th energy consumption terminal in the (k+n)-th control cycle. This represents the in-transit heating and cooling contribution generated by historical control commands during the (k+n)th control cycle. The total number of heat and cold sources. This is the offset from the historical control cycle relative to the current control cycle. To cover the historical control cycles within the effective time range of the corresponding response kernel, The duration of a single control cycle. For the i-th heat source in the i-th... Each control cycle input corresponds to the actual cooling and heating power of the energy supply path. To correspond to the historical hot and cold power after (n+ The response kernel function value applied to the j-th energy-consuming terminal after time Δ.

8. The intelligent control method for distributed data center cold and heat sources based on multimodal perception according to claim 7, characterized in that, The specific steps for determining the distribution of cooling and heating power from each cooling and heating source room to each energy-consuming terminal and the source-side action sequence that matches the demand of the energy-consuming terminals include: Using the cooling and heating power allocated by the i-th heat source to the j-th energy-consuming terminal in the future control cycle as the decision quantity, the expected cooling and heating power to be received at the terminal is calculated according to the following formula: ； In the formula, Let j be the expected cooling or heating power to be received by the j-th energy-consuming terminal in the (k+n)-th control cycle. This represents the offset of future control commands relative to the current control cycle. Let i be the heating or cooling power that the i-th heating or cooling source is intended to allocate to the j-th energy-consuming terminal in the k+ρ control cycle. The proposed allocation of heating and cooling power is processed through (n) The proportion of energy reaching the j-th energy-consuming terminal after time ρ)Δ; Under the conditions of satisfying the limits on the deviation of cold and hot power at the end of the energy consumption terminal, the upper and lower limits of the output of cold and hot source equipment, the rate of change of output of cold and hot source equipment, the shortest continuous operating time, the allowable flow rate of the water pump, the range of water supply temperature, and the mutual exclusion constraints of cold and hot conditions in the same hydraulic branch, the cold and hot power allocation result with the lowest input energy consumption shall be determined according to the following formula: ； In the formula, Let i be the measured input power function of the i-th heat source and cold source. Indicates that all time domains are controlled by the future. The set of decision variables to be optimized consists of J, which is the total number of energy-consuming terminals, and N, which is the number of control cycles contained in the future control time domain.

9. The intelligent control method for distributed data center cold and heat sources based on multimodal perception according to claim 8, characterized in that, The steps of generating corresponding control commands and issuing the control commands at the corresponding command issuance time specifically include: The output of the cold and heat source equipment is determined by the sum of the cold and heat power distributed from the same cold and heat source to each energy-consuming terminal. The target mass flow rate is determined based on the output of the cold and heat source equipment, the temperature difference between the supply and return water, and the specific heat capacity of the transport medium. The target mass flow rate is converted into the pump operating frequency based on the pump performance curve. The opening degree of the branch regulating valve is determined based on the flow characteristics of the regulating valve and the pressure difference of the power supply path. The command issuance time for cold and heat source equipment, water pumps, and branch regulating valves is determined by subtracting the transmission delay of the corresponding energy supply path from the time of cold and heat demand at the end of the energy consumption terminal.

Images