A water-cooled air conditioning system control method suitable for overtime work

Abstract

The present invention relates to the field of intelligent refrigeration control technology, specifically a control method for a water-cooled air-conditioning system suitable for overtime operation, comprising the following steps: collecting static pressure values through a pressure differential sensor to calculate the pressure differential offset, performing sliding average noise reduction and fitting a trend curve, extracting the pressure differential change rate through multi-window analysis, generating a diversion instruction when the temperature difference deviates, adjusting the valve to the target opening and generating a completion signal, gray-predicting the temperature change to generate a failure time window, detecting the flow matching the guide valve opening to generate a compensation signal. In the present invention, trends are extracted through static pressure and pressure differential analysis, combined with sliding average and time series fitting, to dynamically identify the hidden dangers of flow imbalance, adjust the relay valve step diversion, suppress the expansion of non-target branches, use a gray model to predict residual cooling failure, optimize the return path and response efficiency, achieve precise adjustment and temperature compensation of the guide valve, build a closed loop of cooling control, and improve the energy efficiency and stability of the air-conditioning system.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent refrigeration control, and in particular to a control method for a water-cooled air-conditioning system suitable for overtime operation. Background Art

[0002] The field of intelligent refrigeration control technology encompasses dynamic adjustment and optimization of refrigeration system operating conditions, energy consumption management, environmental sensing and regulation, and system linkage management. The core of this technology lies in achieving efficient operation of refrigeration equipment in differentiated usage scenarios through sensor data acquisition, control strategy setting, and actuator response mechanisms. This area encompasses coordinated control of hardware and software, the development of temperature and humidity adjustment strategies, time-based energy management, and the matching of load forecasting with equipment operating cycles. It is commonly used for intelligent operation and scheduling of air conditioning systems in office buildings, industrial plants, and public spaces.

[0003] Among them, the water-cooled air-conditioning system control method suitable for overtime refers to a water-cooled air-conditioning system operation control strategy proposed to address the problem of fluctuations in cooling demand during non-standard working hours in office buildings. The subject of this patent mainly involves collecting regional personnel activity data and indoor temperature change data based on the actual usage of the office area, identifying the air-conditioning load demand during overtime hours, and determining the startup range of the refrigeration equipment based on the time period setting and personnel distribution. It controls the output capacity of the water-cooled air-conditioning system in energy-saving operation mode by dynamically adjusting the operating time of the chilled water pump, frequency conversion control of the cooling tower fan speed, and prioritizing the adjustment of the operating parameters of the high-efficiency refrigeration unit, thereby achieving targeted scheduling control of differentiated usage areas during overtime.

[0004] Existing control models for overtime scenarios often rely on preset time periods and fixed zones, making it difficult to identify actual usage and cooling load fluctuations in real time. This results in delayed control responses. When load fluctuations at the branch end occur suddenly, the system often fails to quickly identify temperature trends and flow rate deviations, resulting in overcooling or uneven heating and cooling in some areas, impacting comfort. Existing systems often rely on single temperature and humidity parameters for decision-making, failing to fully utilize multi-dimensional data such as pressure, flow rate, and occupant distribution. This results in a lack of targeted and forward-looking control decisions. Furthermore, valve adjustments are often performed using static settings, ignoring the dynamic load-adjusted timing requirements for valve opening. This leads to slow response and large execution errors. This is particularly true in office building overtime scenarios, resulting in wasted cooling in non-target areas and insufficient cooling in target areas. For example, when employees on certain floors work overtime, the system still provides cooling in a full-floor mode, increasing the energy burden and affecting local load balance, hindering the precise implementation of cooling control strategies. Summary of the Invention

[0005] To address the problem that existing control modes in overtime scenarios often rely on preset time periods and fixed area settings, making it difficult to identify actual usage status and cooling load fluctuations in real time, resulting in delayed control response. When the load at the end of a branch suddenly changes, the system often cannot quickly identify temperature change trends and flow rate deviation characteristics, resulting in overcooling or uneven heating and cooling in some areas, affecting the comfort experience. Because existing systems often use a single temperature and humidity parameter as the judgment basis, they fail to fully utilize multi-dimensional data such as pressure, flow rate, and personnel distribution, resulting in a lack of targeted and forward-looking control decisions. Furthermore, valve adjustment is often performed using a static setting, ignoring the requirements of dynamic load adjustment for the opening adjustment rhythm. This leads to slow response and large execution errors in the adjustment process. Especially in overtime scenarios in office buildings, it is prone to cooling waste in non-target areas and insufficient cooling supply in target areas. For example, when employees on some floors work overtime, the system still uses a full-floor cooling mode, which not only increases the energy burden, but also affects the local load balance of the system, limiting the effectiveness of the refined implementation of the cooling control strategy. The present invention provides a control method for a water-cooled air conditioning system suitable for overtime. The technical solution is as follows:

[0006] On the one hand, a water-cooled air conditioning system control method suitable for overtime work is provided, the method comprising:

[0007] S1: The static pressure value at the end of the branch is collected through the differential pressure sensor, the static pressure offset of the adjacent branch is calculated, the offset is denoised using the sliding average filter algorithm, and trend fitting is performed according to the time series to generate a pressure differential trend curve;

[0008] S2: performing a multi-window dynamic gradient analysis on the pressure difference trend curve to extract the normalized pressure difference change rate. When steady-state pressure decay is present in consecutive windows and the target branch temperature deviates from the thermal equilibrium reference value, generating a diversion start instruction;

[0009] S3: According to the diversion start instruction, the opening value of the target branch relay valve is identified. If it does not reach the full-open threshold, it is increased to full-open according to the step length. If the opening value of the non-target branch exceeds the locking threshold, it is decreased to locking according to the step length, and a valve adjustment completion signal is generated;

[0010] S4: Based on the valve adjustment completion signal, monitor the branch terminal temperature, collect temperature changes and input them into the grey prediction model, output the time point when the temperature reaches the high temperature threshold, and generate the residual cooling failure time window;

[0011] S5: Detect branch flow based on the residual cooling failure time window, determine the area in combination with the personnel concentration strategy, select branches that meet the load conditions to build a priority return path, calculate the pilot valve opening through the flow offset, and generate a residual cooling compensation execution instruction.

[0012] As a further solution of the present invention, the multi-window dynamic gradient analysis calculates the thermal response characteristics of the branch by analyzing the heat capacity, flow rate, and temperature parameters of the branch, and sets a threshold value for the dynamic gradient;

[0013] The fully open threshold is the upper limit of the valve controllable adjustment range, and the locked threshold is the lower limit of the valve controllable adjustment range. It is set according to the valve adjustable flow rate, and the step size is set based on the degree of opening change within the adjustment cycle;

[0014] The high temperature threshold is dynamically adjusted based on the air conditioning load parameter types and weight coefficients including instantaneous load, original load trend, regional cooling and heating demand ratio, external environmental parameters, etc.;

[0015] The grey prediction model adopts a single variable grey prediction model, the sampling period is set according to the thermal response characteristics of the monitored object, and the data length covers a complete thermal dynamic stage;

[0016] The load condition is that the current flow rate is lower than the design load capacity;

[0017] The pressure difference trend curve includes the pressure difference change amplitude, change rate characteristics, and trend stability index; the diversion start instruction includes the diversion trigger condition type, pressure decay judgment logic, and temperature difference response level; the valve adjustment completion signal includes the relay valve response status, valve action cycle, and adjustment status identifier; the residual cooling failure time window includes the temperature control response time, prediction confidence interval, and heat capacity correction coefficient; the residual cooling compensation execution instruction includes the return path priority, flow distribution ratio, and diversion strategy parameters.

[0018] As a further solution of the present invention, the specific steps of S1 include:

[0019] S101: Collect the static pressure value of the branch at the end of the branch through the pressure difference sensor, calculate the absolute difference between the static pressure values of adjacent branches, store the difference sequence in order by timestamp, and generate the branch static pressure offset;

[0020] S102: calling the branch static pressure offset, applying a sliding average filtering algorithm, performing a convolution operation on the sequence data based on an adaptive time window of branch pressure fluctuation characteristics, and generating a filtered offset sequence;

[0021] The sliding average filtering algorithm is an adaptive sliding filtering algorithm combined with the branch topology structure, and the window length is set according to the static pressure sampling frequency and noise characteristics;

[0022] S103: Based on the filtered offset sequence, a cubic polynomial function relationship is established, trend component values corresponding to time nodes are calculated using the least squares method, and component coordinate points of continuous time intervals are connected to generate a pressure difference trend curve.

[0023] As a further solution of the present invention, the specific steps of S2 include:

[0024] S201: Based on the pressure differential trend curve, construct multiple groups of time sliding windows according to a preset time scale, calculate the pressure differential difference between the start and end times of the window, calculate the change rate based on the difference and the time span, record the corresponding change rate and change direction interval, and generate a normalized pressure differential change rate sequence;

[0025] S202: calling the normalized pressure difference change rate sequence, identifying the continuously decreasing section, extracting the section temperature data, comparing the temperature sequence point by point based on the thermal balance reference value, screening the temperature points whose offset exceeds the reference interval, setting the dynamic temperature offset threshold based on the branch heat capacity parameter, screening the offset abnormal points, and generating the steady-state offset interval identification value;

[0026] The thermal balance reference value is set based on the interval formed by k times the standard deviation above and below the mean value;

[0027] S203: Based on the steady-state offset interval identification value, track the corresponding window index, select the overlapping index set that meets the pressure difference continuous drop and temperature offset conditions, summarize them in index order to form an instruction list, and output the diversion start instruction;

[0028] The condition for the pressure difference to continue to decrease is that a linear fit is performed on the pressure difference value using the least squares method within the sliding window length, and the slope of the fitting line is negative.

[0029] As a further solution of the present invention, the pressure difference change rate is calculated using the formula:

[0030] ;

[0031] in, Representative The pressure difference change rate within a time sliding window is Representative The first time sliding window The pressure difference between the pressure difference value at the moment and the pressure difference value at the start of the window, in Pa, Representative The arithmetic mean of the pressure difference within a time sliding window, in Pa, Representative The first time sliding window The weighting coefficient corresponding to each moment, Representative The time span length of a time sliding window, in seconds, Representative The first time sliding window The time interval between the moment and the start time of the window, in seconds. Representative The number of time points included in a time sliding window.

[0032] As a further solution of the present invention, the specific steps of S3 include:

[0033] S301: Based on the diversion start instruction, obtain the current opening value of the target branch relay valve, compare the current opening value with the set full-open threshold, and if the current opening value is less than the full-open threshold, call the current opening value and the set step size for cumulative calculation, update the real-time opening value of the target branch relay valve, and generate the target branch opening change rate;

[0034] S302: Based on the target branch opening change rate, the current non-target branch relay valve opening value is detected, and the detected opening value is compared with the lockout threshold. If the opening value is higher than the lockout threshold, the current opening value and the set step size are decremented, the opening state of the non-target branch relay valve is updated, and the non-target branch opening change rate is generated;

[0035] S303: calling the target non-branch opening change rate and the target branch opening change rate to determine whether the target branch opening value has reached the full-open threshold and whether the non-target branch opening value is lower than the locking threshold. If both conditions are met, generating a valve adjustment completion signal;

[0036] The step size is calculated as the maximum valve opening×response coefficient / response delay time, where the response coefficient is an empirically determined adjustment sensitivity parameter ranging from 0.1 to 0.5, and the response delay time is in seconds.

[0037] As a further solution of the present invention, the non-target branch opening change rate is calculated using the formula:

[0038] ;

[0039] in, Representative The second adjustment The opening change rate of a non-target branch, in % / step, Representative The second adjustment The current opening value of the non-target branch relay valve, in percentage, Representative The setting step of the non-target branch relay valve is in percentage. Representative Among the non-target branches The adjustment weight coefficient of the influence of each associated valve on the regulation of this branch is obtained through fluid simulation and experimental data fitting based on the physical coupling relationship of the valve. Representative Among the non-target branches The real-time opening value of the associated valve, in percentage, Representative The blocking threshold of non-target branches, in percentage, Representative The average opening value of the valves associated with the non-target branches, in percentage, Represents the number of associated valves in the non-target branch, Represents the current adjustment step number, the unit is step, Indicates the number of the non-target branch. Representative branch Corresponding valve response delay correction factor, unit is 1 / s, Representative branch Valve response delay time, in seconds.

[0040] As a further solution of the present invention, the specific steps of S4 include:

[0041] S401: Based on the valve adjustment completion signal, the terminal temperature of the monitoring branch is monitored, the continuous monitoring time and the corresponding temperature data are extracted, the temperature difference between adjacent moments is calculated to construct a variation sequence, and a temperature variation time series value is obtained;

[0042] S402: Calling the temperature change time series value and inputting it into the grey prediction model, generating a cumulative sequence and performing mean smoothing processing, identifying the time node in the prediction sequence that first exceeds the high temperature threshold, and obtaining the predicted over-temperature time point value;

[0043] S403: Calculate the time offset according to the predicted over-temperature time point value and the monitoring timestamp, set the interval extension width and combine them to form an integrated value of the complete time period to obtain the residual cooling failure time window.

[0044] As a further solution of the present invention, the specific steps of S5 include:

[0045] S501: Detect branch flow based on the residual cooling failure time window, extract branch flow change values per unit time, determine branch carrying capacity, and select branches with controllable fluctuations. Information on overtime worker position distribution and regional concentration expectations is introduced, combined with branch flow change characteristics, to identify branches with long-term stability and adjacent to concentrated areas, and establish a stable flow change interval.

[0046] S502: Based on the branch numbers in the stable flow change interval, the branch location data and traffic load labels are retrieved to determine path connectivity and select branches that meet the load range. In combination with the overtime concentration strategy, the geographical coupling between the available branches and the preset concentration area is calculated. The optimal path group is selected using the cooling response time as a weight, and a load ratio distribution value of the available paths is generated.

[0047] S503: Based on the available path load ratio distribution value and the matching flow offset percentage, the required opening of the pilot valve is calculated and the regulation coefficient is adjusted. Considering the real-time cooling load demand in the concentrated area and the dynamic change trend during overtime hours, the pilot valve opening and regulation priority are corrected in real time to generate a residual cooling compensation execution instruction.

[0048] As a further solution of the present invention, the path load ratio distribution value is calculated using the formula:

[0049] ;

[0050] in, Representative branch With path The available path load ratio distribution value is used as a quantitative indicator for determining the branch return priority, and the pilot valve opening is adjusted to perform dynamic flow guidance. Representative branch The average flow rate change per unit time in the current cycle, in m 3 / s, which is normalized to dimensionless parameters by Z-score. Representative Path The current cycle average flow load adjustment coefficient of the corresponding branch group, Representative branch With path The connectivity judgment parameter between them is 1 if connected and 0 if not connected. Representative branch With path The original average load flow value, in m 3 / s, which is normalized to dimensionless parameters by Z-score. Represents the current cycle branch To Path Total flow rate, in m 3 , which is normalized to a dimensionless parameter by Z-score. Representative branch Corresponding path The maximum number of configured branches, Representative Path Lower branch The standard deviation of the load flow in the current cycle, in m 3 / s.

[0051] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0052] By real-time acquisition of static pressure at the branch end and dynamic analysis of the pressure differential between adjacent branches, and by introducing sliding average filtering and time series trend fitting, it is possible to accurately grasp the pressure change trend. Through multi-window gradient dynamic analysis, the pressure differential change rate is extracted, and combined with the thermal balance benchmark offset, it is determined whether there is a risk of flow imbalance in the branch, thus providing clear trigger conditions for the diversion action. Furthermore, by adjusting the opening of the relay valve and utilizing step-size control to achieve diversion of the target branch, unnecessary opening expansion of non-target branches is suppressed, thus achieving spatial focusing of heat load regulation. In the detection of dynamic temperature changes, the gray prediction model is combined to predict the time point of residual cooling failure in advance, providing a prediction window for cooling capacity scheduling and ensuring timely cooling capacity compensation. Through dynamic branch flow detection and load path optimization, the return flow path is matched and responsive, achieving precise closed-loop adjustment of the pilot valve control under flow offset. The above process realizes a full-process closed loop in terms of pressure difference trend extraction, diversion timing judgment, valve linkage control, temperature prediction compensation and return path optimization, which enhances the timeliness, spatial targeting and execution accuracy of cooling capacity allocation, and improves the energy-saving control capability and operational stability of the air-conditioning system in differentiated usage scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 Schematic diagram of the workflow of the present invention. DETAILED DESCRIPTION

[0054] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0055] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as an "exemplary" in the present invention should not be interpreted as being preferred or advantageous over other embodiments or designs. Rather, the use of the word "exemplary" is intended to present concepts in a concrete manner. Furthermore, in the embodiments of the present invention, "and / or" can mean both or either of the two.

[0056] In the embodiments of the present invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, when the distinction is not emphasized, the meanings they convey are the same. The terms "of," "corresponding," and "corresponding" may sometimes be used interchangeably. It should be noted that, when the distinction is not emphasized, the meanings they convey are the same.

[0057] In the embodiments of the present invention, sometimes a subscript such as W1 may be written as a non-subscript such as W1. When the difference is not emphasized, the meanings to be expressed are the same.

[0058] In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0059] See also Figure 1 The embodiment of the present invention provides a water-cooled air conditioning system control method suitable for overtime work. The processing flow of the method may include the following steps:

[0060] S1: The static pressure value at the end of the branch is collected through the differential pressure sensor, the static pressure offset of the adjacent branch is calculated, the offset is denoised using the sliding average filter algorithm, and trend fitting is performed according to the time series to generate a pressure differential trend curve;

[0061] S2: Perform multi-window dynamic gradient analysis on the pressure difference trend curve to extract the normalized pressure difference change rate. When steady-state pressure decay is present in the continuous window and the target branch temperature deviates from the thermal equilibrium reference value, a diversion start instruction is generated;

[0062] S3: According to the diversion start instruction, the target branch relay valve opening value is identified. If it does not reach the full-open threshold, it is increased to full-open by step. If the non-target branch opening exceeds the locking threshold, it is decreased to locking by step, and a valve adjustment completion signal is generated;

[0063] S4: Based on the valve adjustment completion signal, monitor the branch terminal temperature, collect temperature changes and input them into the grey prediction model. The output is the time point when the temperature reaches the high temperature threshold, and the residual cooling failure time window is generated.

[0064] S5: Detect branch flow based on the residual cooling failure time window, determine the area based on the personnel concentration strategy, select branches that meet the load conditions to build a priority return path, calculate the pilot valve opening through the flow offset, and generate the residual cooling compensation execution instruction.

[0065] Specifically, the steps of S1 are:

[0066] S101: Collect the static pressure value of the branch at the end of the branch through the pressure difference sensor, calculate the absolute difference between the static pressure values of adjacent branches, store the difference sequence in order by timestamp, and generate the branch static pressure offset;

[0067] The process of collecting the static pressure value of the branch end through the pressure differential sensor needs to be configured in combination with the branch distribution in the specific building water supply system. First, the terminal monitoring point location in the water supply branch network is selected, and digital pressure differential sensors are installed at multiple locations. For example, pressure monitoring points are arranged at the ends of the water supply branches on the six floors of the building, and the corresponding numbers are P1 to P6. The sampling frequency is set to once per minute. At each time point, the static pressure values P1(t) to P6(t) of the six branch ends are obtained respectively. They are used as a set of static pressure data to complete a complete data collection at time t. Then, for Two adjacent branches, such as P1 and P2, P2 and P3, and so on, perform the static pressure difference calculation operation between the adjacent branches, that is, perform the difference between the two in absolute value mode, for example, P2(t)-P1(t) takes the absolute value as ΔP1(t), and the same method can be used to calculate ΔP2(t), ΔP3(t)...ΔP5(t). The above differences constitute a set of difference sequence {ΔPi(t)}; this difference sequence will form a key-value pair with the timestamp t and be stored in the database in ascending order of sampling time. In the case of sampling once per minute, 60 will be formed within 1 hour. For 24-hour periodic sampling, 1440 sets of difference data are recorded and stored in the "branch pressure difference database" through unified numbering and timestamp correspondence; in an actual engineering project, taking the water supply system on the 6th floor of an office building as an example, assuming that the data collected at a certain time point t=10:00 are P1=0.42MPa, P2=0.44MPa, P3=0.45MPa, P4=0.43MPa, P5=0.41MPa, P6=0.39MPa, then the pressure differences of adjacent branches are ΔP1=0.02MPa, ΔP2=0. .01MPa, ΔP3=0.02MPa, ΔP4=0.02MPa, ΔP5=0.02MPa, and the corresponding records are stored in the database; the specific execution of the above "calculation" action is: subtract P(i)(t) from P(i+1)(t) and take its absolute value. The process cannot be simplified to a function call and must perform specific numerical operations. For example, if P(i+1)=0.44, P(i)=0.42, the calculation formula is |0.44-0.42|=0.02. When "judging" whether there is a branch with significant pressure difference deviation, it is necessary to set the static pressure difference reference threshold ΔP th =0.03MPa, if ΔP(i)(t)≥ΔP th The branch is judged to have abnormal deviation; the setting of the threshold refers to the daily average pressure difference variation range under the normal operation of the branch. Combined with the 6-month monitoring results, the maximum intraday fluctuation value is 0.027MPa. In order to ensure the analysis sensitivity, ΔP is set. th 0.03MPa, if it exceeds, the deviation judgment will be triggered; in the above sampling data, ΔP(i)(t)<ΔP th, so no abnormal deviation was found at this moment; Table 1 shows the branch static pressure and pressure difference values at different time nodes on a certain day.

[0068] Table 1: Branch static pressure and pressure difference monitoring table (unit: MPa)

[0069]

[0070] As shown in Table 1, the ΔP value at the differentiation moment fluctuates, but the operation is considered stable before it exceeds 0.03 MPa. If the frequency of ΔP3 or ΔP4 reaching or exceeding 0.03 MPa exceeds 5 times per hour, it will be considered that a branch has a deviation trend. By sorting and recording the time series, a complete deviation trend data sequence can be obtained, laying the foundation for subsequent filtering processing.

[0071] S102: Calling the branch static pressure offset, applying a sliding average filtering algorithm, performing a convolution operation on the sequence data based on an adaptive time window of the branch pressure fluctuation characteristics, and generating a filtered offset sequence;

[0072] The sliding average filter algorithm is an adaptive sliding filter algorithm combined with the branch topology structure, and the window length is set according to the static pressure sampling frequency and noise characteristics;

[0073] Based on the branch static pressure offset, the raw data sequence is processed using a sliding average filtering algorithm. In practice, the collected static pressure difference sequence must first be input into the filtering algorithm, and an appropriate sliding window must be selected. The length of this sliding window depends on the sampling frequency and noise characteristics of the static pressure data. For example, in actual applications, if the sampling frequency is set to once per minute, the length of the sliding window can be adjusted based on historical data fluctuations. Assuming the initial setting is 5 minutes, the filtering window contains 5 sampling points. The goal of the sliding average filter is to reduce abnormal data fluctuations caused by environmental noise, making the data more stable and facilitating subsequent trend analysis.

[0074] The specific implementation process is as follows. Assume that at a certain time t, the difference sequence from ΔP1(t) to ΔP5(t) is [0.02, 0.01, 0.02, 0.02, 0.02]. First, a sliding window process is performed on these data with a window size of 5. The goal is to average the current data point with the previous 4 data points to obtain a smooth value. For example, at t=10:00, the first step is to calculate ΔP avg=(0.02+0.01+0.02+0.02+0.02) / 5=0.018MPa; the sliding window then continues moving forward, each time calculating the average of the current position data and the previous four data points within the window to obtain a smoothed ΔP sequence. In this process, the length of the sliding window is a critical parameter, depending not only on the sampling frequency but also on the actual pressure fluctuation characteristics. If excessive noise is detected within a certain period during data analysis, the window length can be increased; otherwise, it can be reduced.

[0075] Once the static pressure difference series has been smoothed, the data can be further optimized using an adaptive sliding filter algorithm. The filter window is adaptively adjusted based on data fluctuations, allowing the window size to change dynamically. For example, during periods of relatively stable operation, the window size can be reduced to reduce unnecessary calculations. During periods of significant pressure fluctuations, the window size automatically increases to better track data changes and eliminate the potential for transient outliers to interfere with trend analysis.

[0076] S103: Based on the filtered offset sequence, a cubic polynomial function relationship is established, trend component values corresponding to time nodes are calculated using the least squares method, and component coordinate points of consecutive time intervals are connected to generate a pressure difference trend curve;

[0077] Based on the filtered data, a trend curve is constructed to reflect the long-term trend of the pressure difference. To this end, a cubic polynomial function is used to establish the mathematical relationship between the pressure difference and time, and the trend component corresponding to the time node is calculated using the least squares method. The core idea of this method is to fit a smooth curve to describe the main trend changes in the data series, eliminating the influence of local noise and short-term fluctuations.

[0078] The specific execution steps are as follows. Assume that the ΔP sequence after sliding average filtering is [0.02, 0.01, 0.02, 0.02, 0.02]. First, use time t as the independent variable and ΔP as the dependent variable to construct a polynomial equation. For example, suppose we choose a cubic polynomial to fit the data. The standard form of the cubic polynomial is:

[0079] ,

[0080] Next, the coefficients a3, a2, a1, and a0 of this equation are calculated using the least squares method. The least squares method involves substituting the time point t and the corresponding ΔP value into the polynomial equation and using an optimization algorithm to minimize the residual error at that point. This involves adjusting a3, a2, a1, and a0 to make the fitted curve most similar to the actual data. Assuming the ΔP sequence [0.02, 0.01, 0.02, 0.02, 0.02] and the corresponding time points t = [1, 2, 3, 4, 5], we substitute these data into the polynomial equation and solve it using the least squares method to obtain the fitted polynomial function.

[0081] By solving the obtained coefficients, the fitted polynomial equation is obtained as follows:

[0082] ;

[0083] Based on this equation, we can further predict the ΔP value at future time points. For example, the predicted ΔP value at t=6 is:

[0084] ;

[0085] Connect the calculated trend component values of the time nodes to generate a complete pressure difference trend curve.

[0086] Specifically, the steps of S2 are:

[0087] S201: Based on the pressure differential trend curve, construct multiple time sliding windows according to a preset time scale, calculate the pressure differential difference between the start and end times of the window, calculate the change rate based on the difference and the time span, record the corresponding change rate and change direction interval, and generate a normalized pressure differential change rate sequence;

[0088] Based on the pressure difference trend curve, we first select the historical pressure difference data sequence of a heating branch. The data is sampled at a frequency of 1 minute and recorded as a sequence , in the time scale Within minutes, a sliding window group is constructed with a sliding interval of 5 minutes. The length of each window is 20 minutes. For each sliding window , record the total Sampling points, and determine the window starting pressure difference value , for any Sampling points, the pressure difference For example, if the initial pressure difference of a window is 162.4kPa and the pressure difference of the fifth point in the window is 161.6kPa, then the corresponding pressure difference is kPa, and then weight the difference in the entire window. In order to reduce human intervention, equal weight processing is adopted, and the weight coefficient of each sampling point is set to , the average value of the difference within the window is , for example: if the difference in the window is There are 20 points in total, and their mean is assumed to be kPa, and then calculate the weighted sum of the weighted offsets, , and then find the weighted time interval and , assuming the sampling time interval is 60 seconds, The time interval between the starting point and the , for example: the 5th point seconds, so a window The weighted sum is:

[0089] ;

[0090] To calculate the rate of change of differential pressure, use the formula:

[0091] ;

[0092] in, Representative Normalized pressure difference change rate within a time sliding window, Representative The first time sliding window The pressure difference between the pressure difference value at the moment and the pressure difference value at the start of the window, in Pa, Representative The arithmetic mean of the pressure difference within a time sliding window, in Pa, Representative The first time sliding window The weighting coefficient corresponding to each moment, Representative The time span length of a time sliding window, in seconds, Representative The first time sliding window The time interval between the moment and the start time of the window, in seconds. Representative The number of time points included in a time sliding window.

[0093] Substitute into the calculation:

[0094] ;

[0095] The result here is 0, which means that the pressure difference in this window is decreasing uniformly. ,but:

[0096] ;

[0097] Continue to calculate the sliding window as above , get the original pressure difference change rate sequence , normalize it and set the overall maximum value to , the minimum value is , then the normalized pressure difference change rate of a certain window is:

[0098] ;

[0099] If the normalized result is less than 0.5, it is marked as "decline" and combined with the start time of the window and end time , record the results and form the content shown in the following table.

[0100] Table 2: Normalized pressure difference change rate sequence table

[0101]

[0102] As shown in Table 2, the normalized change rates within the differentiated time windows are all positive, but the corresponding original change rates are in the negative direction, so they are uniformly marked as "declining". The field values in the table are directly calculated through the measured pressure difference and the time series. The normalization process uses a linear interpolation method to make the window change rate within the standard range of [0, 1].

[0103] S202: Calling the normalized pressure difference change rate sequence, identifying the continuously decreasing section, extracting the section temperature data, comparing the temperature sequence point by point based on the thermal balance reference value, screening the temperature points whose offset exceeds the reference interval, setting the dynamic temperature offset threshold based on the branch heat capacity parameter, screening the offset abnormal points, and generating the steady-state offset interval identification value;

[0104] The thermal balance reference value is set based on the interval formed by k times the standard deviation above and below the mean value;

[0105] After calling the normalized pressure difference change rate sequence [0.11, 0.20, 0.34], the change rate direction values of adjacent sliding windows are compared in sequence [-1, -1, -1] to determine whether there is a continuous negative area. Since the direction here is all negative, the period [00:00, 00:30] is judged to be a continuously decreasing segment. Extract the temperature sequence [65.2, 64.8, 64.5, 63.9, 63.5, 63.0] within the corresponding time segment to set the thermal balance reference interval. Calculate the mean of this sequence , standard deviation .use Set the thermal equilibrium range to The temperature series was compared point by point to see if it fell within the interval. It was found that all data points were within the interval, and it was preliminarily judged that there was no obvious deviation.

[0106] Further Set to 1, the thermal equilibrium range is adjusted to At this time, the temperature values 63.0, 63.5, and 63.9 in the sequence are lower than the lower limit of the interval or close to the lower limit, and the offset point is initially screened. The heat capacity is introduced to judge the actual offset degree, and the branch heat capacity is set , temperature offset impact threshold , and the temperature offset threshold is obtained .by As a benchmark, screen to meet The temperature points are retained, that is, the temperature values less than 63.55 or greater than 64.75 are retained. The values 63.0 and 63.5 meet the conditions and correspond to the times 00:25 and 00:30, respectively, and are marked as steady-state offset intervals. The final offset interval identification value array is [0, 0, 0, 0, 1, 1].

[0107] S203: Based on the steady-state offset interval identification value, track the corresponding window index, select the overlapping index set that meets the pressure difference continuous drop and temperature offset conditions, summarize them in index order to form an instruction list, and output the diversion start instruction;

[0108] The condition for the pressure difference to continue to decrease is that the pressure difference value is linearly fitted using the least squares method within the sliding window length, and the slope of the fitting line is negative;

[0109] Based on the steady-state offset interval identifier value [0, 0, 0, 0, 1, 1] and the normalized pressure differential change direction value [-1, -1, -1], the index positions corresponding to the intersection of the two are extracted as the last two items in the sequence, with indices 4 and 5, corresponding to the times [00:25, 00:30]. This identifies the overlapping segments that meet the dual conditions of "continuous pressure differential decrease" and "stable temperature offset," and summarizes them to generate a list of diversion start instructions [4, 5]. Based on the sliding window start time pointed to by the index, the final trigger time set for the diversion start action is {00:25, 00:30}, which serves as the execution instruction node for the control system to respond to abnormal temperature offsets and pressure differential trends.

[0110] Specifically, the steps of S3 are:

[0111] S301: Based on the diversion start instruction, the current opening value of the target branch relay valve is obtained, and the current opening value is compared with the set full-opening threshold. If the current opening value is less than the full-opening threshold, the current opening value is added to the set step size for accumulation operation to update the real-time opening value of the target branch relay valve and generate the target branch opening change rate;

[0112] After receiving the diversion start command, the digital valve positioner installed on the target branch B1 reads the current opening value θ=72%, and calls the preset full-open threshold θ max=85%, 72% is numerically compared with 85%. When 72% < 85%, the set step size δ is calculated as 100% × 0.3 / 5 = 6% (assuming the maximum valve opening is 100%, the response coefficient β = 0.3, and the response delay time T = 5 seconds). The valve opening accumulation operation is performed: 72% + 6% = 78%, the valve opening value is updated to 78%, and the opening change rate Δθ = 6% / step is calculated. For example, in a chemical reactor feeding system, when the main reaction channel needs to increase the flow rate, the regulating valve is gradually opened in steps of 6%. After three adjustments, the opening value reaches 90% (72% + 3 × 6%). At this time, the change rate remains constant at 6% / step until the full opening threshold is reached. If the opening value of 94% exceeds 85% during the fourth adjustment, the accumulation operation is stopped to generate the final opening change rate.

[0113] S302: Based on the target branch opening change rate, the current non-target branch relay valve opening value is detected, and the detected opening value is compared with the lockout threshold. If the opening value is higher than the lockout threshold, the current opening value and the set step size are decremented, the opening state of the non-target branch relay valve is updated, and the non-target branch opening change rate is generated;

[0114] According to the target branch opening change rate, the non-target branch relay valve opening value is detected. When the current opening value of branch S2 is detected to be 65% ( ), set the lockout threshold to 10% ( ), the real-time opening values of valves V1 and V2 associated with branch S2 are collected through pressure sensors and are 45% and 35% respectively ( , ), the adjustment weight coefficient is determined according to the fluid simulation experimental data 、 , calculate the weighted sum of associated valves: , get the average opening value of branch S2 ( ) = , the response delay time is measured by the timer seconds, set the delay correction factor according to the valve specification manual , calculate the denominator: , and finally calculate the opening change rate , compare the calculated result with the locking threshold. When the rate of change exceeds the locking threshold, the step size reduction operation is performed. For example, the initial opening of 65% is reduced to 50% after 3 adjustments ( ), generate the non-target branch opening change rate, using the formula:

[0115] ;

[0116] in, Representative The second adjustment The opening change rate of a non-target branch, in % / step, Representative The second adjustment The current opening value of the non-target branch relay valve, in percentage, Representative The setting step of the non-target branch relay valve is in percentage. Representative Among the non-target branches The adjustment weight coefficient of the influence of each associated valve on the regulation of this branch is obtained through fluid simulation and experimental data fitting based on the physical coupling relationship of the valve. Representative Among the non-target branches The real-time opening value of the associated valve, in percentage, Representative The blocking threshold of non-target branches, in percentage, Representative The average opening value of the valves associated with the non-target branches, in percentage, Represents the number of associated valves in the non-target branch, Represents the current adjustment step number, the unit is step, Indicates the number of the non-target branch. Representative branch Corresponding valve response delay correction factor, unit is 1 / s, Representative branch Valve response delay time, in seconds.

[0117] calculate ;

[0118] The results show that the opening of the non-target branch S2 needs to be reduced by 3.2% each time it is adjusted. After three adjustments, the opening value is reduced from 65% to 50% ( ), when the opening value is detected to be 10% lower than the locking threshold, the adjustment is stopped. The formula is introduced by the weight coefficient Accurately reflects the multi-valve coupling effect, combined with the delay correction factor Improve response accuracy and ensure synchronous opening adjustment within a 2-second delay time.

[0119] S303: Call the target non-branch opening change rate and the target branch opening change rate to determine whether the target branch opening value has reached the full-open threshold and whether the non-target branch opening value is lower than the locking threshold. If both conditions are met, generate a valve adjustment completion signal;

[0120] The step size is calculated as the maximum valve opening × response coefficient / response delay time. The response coefficient is an empirically determined adjustment sensitivity parameter ranging from 0.1 to 0.5. The response delay time is in seconds.

[0121] Call the target branch B1 opening change rate of 6% / step and the non-target branch S2 change rate of 3.2 / step, and continuously monitor the current opening value θ of the target branch target =90%, and the full-open threshold θ max =85% for numerical comparison, when 90% ≥ 85% is determined to be up to standard, and at the same time the opening value θ of the non-target branch S2 is detected non-target = 8%, and the latching threshold θ min =10%, when 8%≤10%, it is determined to be up to standard. In the heat pipe network balance regulation, when the main heating branch reaches 90% opening after 5 adjustments (initial 60%+5×6%), the non-target branch is reduced to 8% after 3 adjustments (initial 65%-3×5%×3.2%), and θ is satisfied at the same time. target ≥θ max And θ non-target ≤θ min Two conditions trigger the DCS system to output a 4-20mA valve adjustment completion signal.

[0122] Specifically, the steps of S4 are:

[0123] S401: Monitoring the branch terminal temperature based on the valve adjustment completion signal, extracting the continuous monitoring time and corresponding temperature data, calculating the temperature difference between adjacent moments to construct a variation sequence, and obtaining a temperature variation time series value;

[0124] Based on the valve adjustment completion signal, the terminal temperature of the branch is monitored. First, the status feedback of the valve actuator in the temperature control system is received through the signal processing unit. When the feedback signal is marked as the "adjustment completed" state, the terminal temperature of the branch is collected in real time. The temperature data acquisition module sets the sampling period to 60 seconds. During the continuous monitoring process, multiple groups of time and corresponding temperature value combinations are extracted. Each group contains a timestamp and the temperature data at the corresponding moment. For example, in the actual overtime operation environment, the monitoring point is located at the end of the air-conditioning branch on the west side of the third floor of the office building. A total of 5 data points are collected within the monitoring period of 5 minutes; the numerical difference between two consecutive temperature values is calculated, and the difference processing unit uses the direct difference method to calculate the temperature change at adjacent moments, that is, the temperature value at the latter moment is subtracted from the temperature value at the previous moment to form a temperature change sequence. For example, the monitoring data are 27.5℃, 2 8.0℃, 28.6℃, 29.3℃, 30.1℃, and the corresponding changes are 0.5℃, 0.6℃, 0.7℃, and 0.8℃, respectively. Each difference is recorded and combined into a time series array for further temperature change trend extraction. In the above process, to ensure that the difference reflects the true change trend, the temperature data of the monitoring points need to be preliminarily smoothed and abnormal fluctuation values ​​needed to be eliminated. The abnormal identification condition is the record of the change greater than 2℃ / min or less than -2℃ / min. After identification, the data in this section is eliminated and the difference is recalculated. If the data is normal, the constructed temperature change time series is [0.5, 0.6, 0.7, 0.8], which is used as the basis for temperature dynamic judgment in subsequent steps. In order to show the temperature behavior of the differentiated change stage, the above sequence is stored in vector form, and the data frame structure is constructed and input into the subsequent processing module, as shown in Table 3:

[0125] Table 3: Branch end temperature change table

[0126]

[0127] As shown in Table 3, the five sets of temperature data show a steady upward trend in continuous changes. The resulting change sequences are filed into the database one by one, and the sequences are used as important basic data for the dynamic evolution of the temperature state and input into the next stage of the processing flow, ultimately obtaining the temperature change time series value.

[0128] S402: Calling the temperature change time series value and inputting it into the grey prediction model, generating a cumulative sequence and performing mean smoothing processing, identifying the time node in the prediction sequence that first exceeds the high temperature threshold, and obtaining the predicted over-temperature time point value;

[0129] The temperature change time series value is called to input the grey prediction model. First, the temperature change sequence [0.5, 0.6, 0.7, 0.8] constructed in S401 is read through the data cache module. Then, the sequence is accumulated in sequence. The first item is 0.5. The second item is 1.1 obtained by adding 0.5 and 0.6. The third item is 1.1 plus 0.7 to get 1.8. The fourth item is 2.6, thus constructing the accumulated sequence [0.5, 1.1, 1.8, 2.6]. Then, the three-point sliding average method is used to smooth the accumulated sequence. The average values of the previous item, the current item, and the next item of the intermediate data items 1.1 and 1.8 are calculated respectively, which are (0.5+1. 1+1.8) / 3=1.13 and (1.1+1.8+2.6) / 3=1.83, and the boundary items 0.5 and 2.6 are respectively averaged by two points, that is, (0.5+1.1) / 2=0.55 and (1.8+2.6) / 2=2.2, and finally the smoothed sequence [0.55, 1.13, 1.83, 2.2] is obtained. This sequence is used as the basic data for model prediction and input into the next step of deduction. Based on this sequence, a time axis with a time step of 60 seconds is established, and an extended prediction operation is performed. It is predicted that the cumulative temperature changes at the next three time nodes will be 2.9, 3.6, and 4.5, respectively, corresponding to the prediction point time of 300 seconds, 360 seconds, and 420 seconds. On this basis, it is judged whether it has entered the high temperature warning area, and the current actual temperature is converted to the actual temperature. The terminal temperature value of 30.1°C is used as the reference temperature. It is sequentially added to the predicted change value to obtain predicted temperature values of 33.0°C, 33.7°C, and 34.6°C, respectively. The high temperature threshold is set to 30.5°C. This threshold is set as follows: In an overtime simulation experiment, the initial indoor temperature is 27.5°C, the temperature rise rate is 0.8°C / min, and after 4 minutes (240 seconds), the predicted temperature is 27.5 + 0.8 × 4 = 30.7°C. Based on the investigation of the upper limit of the temperature perception threshold in overtime work environments, 30.5°C, slightly lower than this point, is selected as the risk threshold with a safer boundary. When 30.5°C is compared with the predicted temperature results one by one, the first limit crossing occurs at the predicted time point 300 seconds, that is, 33.0°C first exceeds 30.7°C. 5℃, and thus the predicted over-temperature time point is 300 seconds; the thermal dynamic stage experiment based on this step is as follows: under the operating conditions of the terminal air-conditioning system in the office building, the experimental design is to simulate the actual overtime environment. The cooling port is closed 60 seconds after the air-conditioning adjustment is completed, and the temperature reaction process is recorded. The sampling period is 60 seconds, and 8 sets of temperature data are collected within the range of 420 seconds. The temperature curve shows three typical stages: 0~120 seconds The temperature rises slowly as the initial stable stage, and the average temperature change is less than 0.3℃. From 120 to 300 seconds, it enters the linear temperature rise stage, with an average change of 0.6℃ and a maximum of 0.8℃. After 300 seconds, it is a stable over-temperature stage, which is significantly higher than the preset threshold; therefore, the extracted temperature change sequence [0.5, 0.6, 0.7, 0.8] is in the most representative linear temperature rise stage, has thermal response sensitivity, can reflect the actual indoor temperature rise trend, and is suitable as prediction model training data, thus ensuring that the model can accurately identify the predicted overtemperature time point value of 300 seconds.

[0130] S403: Calculate the time offset based on the predicted overtemperature time point value and the monitoring timestamp, set the interval extension width and combine them to form an integrated value of the complete time period to obtain the residual cooling failure time window;

[0131] The time offset is calculated based on the predicted over-temperature time point value and the monitoring timestamp. The current temperature monitoring end timestamp of 240 seconds is used as the reference point, and the predicted over-temperature time point value of 300 seconds is obtained. The offset Δt=60 seconds is calculated. This offset is superimposed and extended with the expected hysteresis of the threshold trigger. In order to construct the actual control window, the interval extension width is set. In this example, the window is set to extend forward by 60 seconds and backward by 120 seconds. Therefore, the integrated value of the complete time period is extended from 240 seconds to 420 seconds. Combined with this window time period, it is integrated into the residual cooling failure time window [240s, 420s]. In the actual overtime scenario, if the air-conditioning branch at the end of the office building fails to achieve the cooling effect during this window period, the compensation mechanism or alarm module is automatically triggered. The above offset calculation The process is as follows: the predicted time point value 300s minus the last monitoring time point 240s is used to obtain an offset Δt = 60s. This offset is used to evaluate the predicted advance response time. To ensure that the control strategy responds promptly, a forward 60s compensation period is set to cope with short-cycle errors, and a backward 120s is set as the thermal inertia cooling capacity decay window, forming a time period of [240s, 420s]. The rationality of the time window is actually evaluated. If the temperature rise rate is 0.8°C per minute, the temperature rise within the window is expected to be 2.4°C within 3 minutes. The temperature rise is calculated from the current temperature of 30.1°C to a maximum of approximately 32.5°C, which is significantly beyond the indoor comfort limit. This verifies that the window setting has room for temperature rise response, and finally constructs the residual cooling failure time window of [240s, 420s].

[0132] Specifically, the steps of S5 are:

[0133] S501: Detect branch flow based on the residual cooling failure time window, extract the branch flow change value per unit time, determine the branch carrying capacity, and select branches with controllable fluctuations. Incorporating information on the distribution of overtime workers and expected regional concentration, combined with branch flow change characteristics, identify branches with long-term stability and proximity to concentrated areas, and establish a stable flow change interval.

[0134] After determining that the residual cooling has failed, the branch flow stability screening phase begins. The residual cooling failure detection starting point is set to 180 seconds after failure, and sampling is performed every 60 seconds to continuously collect instantaneous flow data within three time periods. Taking branch numbers 3, 7, and 11 as an example, the sampling values are: branch 3 is 0.13, 0.14, and 0.12m 3 / min, branch 7 is 0.17, 0.16, 0.18m 3 / min, branch 11 is 0.12, 0.13, 0.11m 3 / min. First calculate the flow change sequence of the branch, for example, the flow change of branch 3 is +0.01 and -0.02m 3 / min, branch 7 changes to -0.01 and +0.02m 3 / min, branch 11 is +0.01 and -0.02m 3 / min, and the branch change value is obtained in turn. The judgment standard is that the absolute value of two adjacent changes is less than 0.05m 3 / min, the fluctuation is considered controllable and is included in the candidate branch set. After initial screening, branches 3, 7, and 11 meet this condition and are included in the candidate set.

[0135] Subsequently, the flow stability was further determined by the standard deviation, and the upper limit standard deviation was set to 0.02m 3 / min, calculate the standard deviation of the branch flow change value. The standard deviation of branch 3 is about 0.015m 3 / min, branch 7 is 0.015m 3 / min, branch 11 is 0.015m 3 / min, all three are below the set threshold and are therefore identified as stable branches. The final output is a stable interval branch number sequence [3, 7, 11]. Combined with the spatial clustering characteristics of the overtime area, the spatial locations of these branches are associated with the proximity of the concentrated area, which serves as the input basis for the subsequent path load ratio and pilot valve control module.

[0136] S502: Based on the branch numbers in the stable flow variation interval, the branch location data and traffic load labels are retrieved to determine path connectivity and select branches that meet the load range. In combination with the overtime worker concentration strategy, the geographical coupling between the available branches and the preset concentration area is calculated. The optimal path group is selected using the cooling response time as a weight, and a distribution value of the available path load ratio is generated.

[0137] According to the branch number in the stable interval of flow change, the branch number sequence [3, 7, 11] in the stable interval is first processed, and the corresponding physical location data is read in sequence to form a coordinate array [(12.3, 45.6), (18.2, 50.4), (21.9, 48.1)]. Each branch number is used as a query index to extract its spatial coordinate value in the location database. After the coordinate extraction is completed, it is recorded in the form of a two-dimensional vector to prepare for the subsequent spatial connectivity judgment. Subsequently, the mean unit time flow change of each branch in the past 5 minutes is extracted through the flow label interface, which is 0.13, 0.17, and 0.12m respectively. 3 / min, the corresponding numbers are 3, 7, and 11, and their flow values are compared with the set maximum carrying flow thresholds one by one. The threshold of branch 3 is 0.25m 3 / min, Branch No. 7 is 0.3m 3 / min, Branch No. 11 is 0.2m 3 / min, calculate the remaining load value between the current flow rate and the threshold and divide it by its maximum load value to obtain the available load ratio of the corresponding branch. The calculation is as follows: The remaining load of branch 3 is 0.25-0.13=0.12m 3 / min, the ratio is 0.12 / 0.25=0.48; branch 7 is 0.3-0.17=0.13m 3 / min, the ratio is 0.13 / 0.3≈0.43; branch 11 is 0.2-0.12=0.08m 3 / min, the ratio is 0.08 / 0.2=0.4, and the minimum available load ratio threshold is set to 30%, that is, the ratio ≥0.3 is determined to be an available branch. After judging that the above branches meet the availability requirements in turn, enter the branch spatial connectivity judgment step, take branch 3 as the reference benchmark, calculate its Euclidean space distance to branch 7 and branch 11, perform vector coordinate difference square operation in turn, sum and then square root processing, calculate the distance from branch 3 to 7 to be 7.61m, and from branch 3 to 11 to be 9.88m, set the connectivity judgment threshold to 10m, if any branch pair is not greater than 10m, it is determined to be spatially connected, branch 3 and branches 7 and 11 all meet this condition, confirming that their spatial connectivity status is "1", and the execution value is δ （3，7） =1,δ （3，11） =1, based on which a connected path branch sequence [3, 7, 11] is generated. At the same time, the spatial overlap of the path group within the overtime area distribution coordinate range is checked and incorporated into the coupling calculation weight to improve the matching efficiency of cooling capacity allocation. Subsequently, the path load ratio values are calculated using the formula:

[0138] ;

[0139] in, Representative branch With path The available path load ratio distribution value is used as a quantitative indicator for determining the branch return priority, and the pilot valve opening is adjusted to perform dynamic flow guidance. Representative branch The average flow rate change per unit time in the current cycle, in m 3 / s, which is normalized to dimensionless parameters by Z-score. Representative Path The current cycle average flow load adjustment coefficient of the corresponding branch group, Representative branch With path The connectivity judgment parameter between them is 1 if connected and 0 if not connected. Representative branch With path The original average load flow value, in m 3 / s, which is normalized to dimensionless parameters by Z-score. Represents the current cycle branch To Path Total flow rate, in m 3 , which is normalized to a dimensionless parameter by Z-score. Representative branch Corresponding path The maximum number of configured branches, Representative Path Lower branch The standard deviation of the load flow in the current cycle, in m 3 / s.

[0140] is the average flow change per unit time of branch i in the current cycle, which is 3.61×10 -3 , 4.72×10 -3 , 3.33×10 -3 , is the average traffic load adjustment coefficient of the current period under path j, which is uniformly set to 0.9 in this example. is 1, is the original average load flow value of the path, and the standardized values are 0.1, 0.15, 0.08, is the total flow from the branch to the path in the current period, which are 0.021, 0.026, and 0.019 m 3 , after standardization, the values are 0.25, 0.30, and 0.22. The number of configured branches under the path is uniformly set to 2. is the standard deviation of the load flow in the current cycle, which is uniformly set to 0.004m3 / s, after conversion, the calculations are as follows:

[0141] ;

[0142] ;

[0143] ;

[0144] The results show that the path load ratio values of branches 3, 7, and 11 are 218.01, 195.16, and 230.27, respectively. Based on these values, a path load ratio array can be generated for subsequent pilot valve regulation and control operations.

[0145] S503: Based on the available path load ratio distribution value and the matching flow offset percentage, the required opening of the pilot valve is calculated and the regulation coefficient is adjusted. Considering the real-time cooling load demand in the concentrated area and the dynamic change trend during overtime hours, the pilot valve opening and regulation priority are corrected in real time to generate the residual cooling compensation execution instruction.

[0146] After obtaining the path load ratio array [218.01, 195.16, 230.27], it is used as the reference indicator of the branch path capacity available in the current cycle. At the same time, the total amount of residual cooling compensation flow adjustment is set to 0.3m 3 / min, the offset ratio is 20%, and the absolute flow offset that needs to be redistributed is calculated to be 0.06m 3 / min. Based on the path load ratio, the proportional distribution is carried out, and the regulated flow rate allocated to branch 3 is about 0.022m 3 / min, the branch is divided into 7 parts, which is about 0.020m 3 / min, branch 11 is about 0.018m 3 This allocation result is used to generate the branch regulation target, and the required regulation increment is calculated based on the current flow status.

[0147] Next, call the pilot valve control module to convert the branch adjustment amount into the pilot valve opening increment. Set each pilot valve to correspond to 0.01m 3 For a flow rate increase of 100 / min, branch 3 needs to be adjusted by approximately 2.2 gears, branch 7 by approximately 2.0 gears, and branch 11 by approximately 1.8 gears. The pilot valve gear values are rounded to the nearest integer and all set to gear 2. Branches near high-incidence overtime work areas are prioritized for adjustment values. The pilot valve control strategy is then prioritized accordingly to ensure priority response to the dynamic needs of concentrated cooling load areas. This ultimately results in a compensation control instruction set of {3:2, 7:2, 11:2}. The control module issues execution commands to adjust the physical opening of the branch pilot valves, thereby completing the actual flow redistribution process for residual cooling compensation during this cycle.

[0148] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A water-cooled air conditioning system control method suitable for overtime work, characterized in that: The following steps are involved: S1: The static pressure value at the end of the branch is collected through the differential pressure sensor, the static pressure offset of the adjacent branch is calculated, the offset is denoised using the sliding average filter algorithm, and trend fitting is performed according to the time series to generate a pressure differential trend curve; S2: performing a multi-window dynamic gradient analysis on the pressure difference trend curve to extract the normalized pressure difference change rate. When steady-state pressure decay is present in consecutive windows and the target branch temperature deviates from the thermal equilibrium reference value, generating a diversion start instruction; S3: According to the diversion start instruction, the opening value of the target branch relay valve is identified. If it does not reach the full-open threshold, it is increased to full-open according to the step length. If the opening value of the non-target branch exceeds the locking threshold, it is decreased to locking according to the step length, and a valve adjustment completion signal is generated; S4: Based on the valve adjustment completion signal, monitor the branch terminal temperature, collect temperature changes and input them into the grey prediction model, output the time point when the temperature reaches the high temperature threshold, and generate the residual cooling failure time window; S5: Detect branch flow based on the residual cooling failure time window, determine the area in combination with the personnel concentration strategy, select branches that meet the load conditions to build a priority return path, calculate the pilot valve opening based on the flow offset, and generate a residual cooling compensation execution instruction; The specific steps of S5 include: S501: Detect branch flow based on the residual cooling failure time window, extract branch flow change values per unit time, determine branch carrying capacity, and screen branches whose fluctuations are within a controllable range. Combined with the spatial clustering characteristics of the overtime area, associate the spatial location of the branch with its proximity to the concentrated area. Combined with the branch flow change characteristics, identify branches with long-term stability and close to the concentrated area, and establish a stable flow change interval. S502: Based on the branch numbers in the stable flow change interval, the branch location data and flow load labels are retrieved to determine path connectivity and select branches that meet the load range. The spatial overlap of the path group within the overtime area distribution coordinate range is checked. The geographical coupling between the available branches and the preset concentration area is calculated. The optimal path group is selected using the cooling response time as a weight, and a distribution value of the available path load ratio is generated. S503: Based on the available path load ratio distribution value and the matching flow offset percentage, the required opening of the pilot valve is calculated and the regulation coefficient is adjusted. Considering the real-time cooling load demand in the concentrated area and the dynamic change trend during the overtime period, the regulation value is preferentially allocated to the branches near the overtime high-incidence area, and the pilot valve control strategy is prioritized accordingly. The pilot valve opening and regulation priority are corrected in real time to generate the residual cooling compensation execution instruction.

2. The water-cooled air-conditioning system control method suitable for overtime work according to claim 1, characterized in that: The fully open threshold is the upper limit of the valve controllable adjustment range, and the locked threshold is the lower limit of the valve controllable adjustment range. It is set according to the valve adjustable flow rate, and the step size is set based on the degree of opening change within the adjustment cycle; The high temperature threshold is dynamically adjusted based on the instantaneous load, original load trend, regional cooling and heating demand ratio, external environmental parameters, air conditioning load parameter type and weight coefficient; The grey prediction model adopts a single variable grey prediction model, the sampling period is set according to the thermal response characteristics of the monitored object, and the data length covers a complete thermal dynamic stage; The load condition is that the current flow rate is lower than the design load capacity; The pressure difference trend curve includes the pressure difference change amplitude, change rate characteristics, and trend stability index; the diversion start instruction includes the diversion trigger condition type, pressure decay judgment logic, and temperature difference response level; the valve adjustment completion signal includes the relay valve response status, valve action cycle, and adjustment status identifier; the residual cooling failure time window includes the temperature control response time, prediction confidence interval, and heat capacity correction coefficient; the residual cooling compensation execution instruction includes the return path priority, flow distribution ratio, and diversion strategy parameters.

3. The water-cooled air conditioning system control method suitable for overtime work according to claim 1, characterized in that: The specific steps of S1 include: S101: Collect the static pressure value of the branch end through the pressure difference sensor, calculate the absolute difference between the static pressure values of adjacent branches, store the difference sequence in order by timestamp, and generate the branch static pressure offset; S102: calling the branch static pressure offset, applying a sliding average filtering algorithm, performing a convolution operation on the sequence data based on an adaptive time window of branch pressure fluctuation characteristics, and generating a filtered offset sequence; The sliding average filtering algorithm is an adaptive sliding filtering algorithm combined with the branch topology structure, and the window length is set according to the static pressure sampling frequency and noise characteristics; S103: Based on the filtered offset sequence, a cubic polynomial function relationship is established, trend component values corresponding to time nodes are calculated using the least squares method, and component coordinate points of continuous time intervals are connected to generate a pressure difference trend curve.

4. The water-cooled air conditioning system control method suitable for overtime work according to claim 3, characterized in that: The specific steps of S2 include: S201: Based on the pressure differential trend curve, construct multiple groups of time sliding windows according to a preset time scale, calculate the pressure differential difference between the start and end times of the window, calculate the change rate based on the difference and the time span, record the corresponding change rate and change direction interval, and generate a normalized pressure differential change rate sequence; S202: calling the normalized pressure difference change rate sequence, identifying the continuously decreasing section, extracting the section temperature data, comparing the temperature sequence point by point based on the thermal balance reference value, screening the temperature points whose offset exceeds the reference interval, setting the dynamic temperature offset threshold based on the branch heat capacity parameter, screening the offset abnormal points, and generating the steady-state offset interval identification value; The thermal balance reference value is set based on the interval formed by k times the standard deviation above and below the mean value; S203: Based on the steady-state offset interval identification value, track the corresponding window index, select the overlapping index set that meets the pressure difference continuous drop and temperature offset conditions, summarize them in index order to form an instruction list, and output the diversion start instruction; The condition for the pressure difference to continue to decrease is that a linear fit is performed on the pressure difference value using the least squares method within the sliding window length, and the slope of the fitting line is negative.

5. The water-cooled air-conditioning system control method suitable for overtime work according to claim 4, characterized in that: The specific steps of S3 include: S301: Based on the diversion start instruction, obtain the current opening value of the target branch relay valve, compare the current opening value with the set full-open threshold, and if the current opening value is less than the full-open threshold, call the current opening value and the set step size for cumulative calculation, update the real-time opening value of the target branch relay valve, and generate the target branch opening change rate; S302: Based on the target branch opening change rate, the current non-target branch relay valve opening value is detected, and the detected opening value is compared with the lockout threshold. If the opening value is higher than the lockout threshold, the current opening value and the set step size are decremented, the opening state of the non-target branch relay valve is updated, and the non-target branch opening change rate is generated; S303: Call the target non-branch opening change rate and the target branch opening change rate to determine whether the target branch opening value has reached the full-open threshold, and determine whether the non-target branch opening value is lower than the locking threshold. If all conditions are met, generate a valve adjustment completion signal.

6. The water-cooled air conditioning system control method suitable for overtime work according to claim 5, characterized in that: The specific steps of S4 include: S401: Based on the valve adjustment completion signal, the terminal temperature of the monitoring branch is monitored, the continuous monitoring time and the corresponding temperature data are extracted, the temperature difference between adjacent moments is calculated to construct a variation sequence, and a temperature variation time series value is obtained; S402: Calling the temperature change time series value and inputting it into the grey prediction model, generating a cumulative sequence and performing mean smoothing processing, identifying the time node in the prediction sequence that first exceeds the high temperature threshold, and obtaining the predicted over-temperature time point value; S403: Calculate the time offset according to the predicted over-temperature time point value and the monitoring timestamp, set the interval extension width and combine them to form an integrated value of the complete time period to obtain the residual cooling failure time window.

Images