A time-division pressure control system and control method

Abstract

The application discloses a kind of time-sharing pressure control system and control method, it belongs to secondary water supply technical field.Time-sharing pressure control system includes: data acquisition unit, obtains original historical data and pump related original data;Water quantity prediction unit, constructs water quantity prediction model according to original historical data and obtains water consumption prediction data;Pressure prediction unit, constructs pressure prediction model according to original historical data and obtains import pressure prediction data;Pressure fitting unit, according to original historical data fitting obtains the first corresponding relationship between export pressure and import and export flow;Pump fitting unit, according to original historical data and pump related original data fitting obtains the second corresponding relationship between export flow and the actual working parameter of water pump;Optimization configuration unit, according to original historical data and the result calculated above carries out time segment segmentation and configures water pump control scheme.Affirmative effect is: time-sharing pressure control is reasonable, and energy consumption is reduced.

Description

Technical Field

[0001] This invention relates to the field of secondary water supply technology, and in particular to a time-segmented pressure control system and control method. Background Technology

[0002] In recent years, with the continuous development of the social economy and the improvement of people's living standards, the demand for water supply has shifted from simply "having water to drink" to demanding high-quality water supply. Traditional residential water supply is a primary supply method, meaning a one-time process of water being directly delivered from a water treatment plant or water source to the user. However, due to historical planning and other factors, the pressure of municipal water supply networks in most cities is limited, leading to unstable pressure in primary supply. Therefore, relying solely on primary supply cannot meet users' demand for "high-quality water." Thus, to achieve high-quality water supply, existing technologies typically employ secondary water supply methods to replace primary supply methods.

[0003] Secondary water supply, compared to primary water supply, refers to the process of treating tap water through an intermediate secondary water supply system before supplying it to users. This secondary system primarily involves filtration, disinfection, softening, and pressurization of the tap water. It effectively removes impurities, bacteria, viruses, and other harmful substances from the tap water and provides sufficient pressure, resulting in more stable water pressure at the user's end. Secondary water supply systems are widely used in residential communities, commercial buildings, industrial parks, and other locations, providing users with a better water experience.

[0004] Traditional secondary water supply systems typically employ a constant pressure system. This system suffers from a pressure setting issue: if the pressure is too low, the supply will be insufficient; conversely, if the pressure is too high, according to the relationship between water leakage (L) and pressure (P) raised to the power of N (L = P^N, where L and N range from 0.5 to 2.5, with an average of 1.15, representing a near-linear relationship), leakage in the pipe network will increase with increasing pressure. Therefore, an increasing number of secondary water supply systems are adopting appropriate pressure control strategies to not only meet supply pressure requirements but also reduce pipe network leakage and lower operating costs.

[0005] However, existing pressure control methods are mostly designed for macro-level pipe networks. For example, various pressure regulation strategies are set up for different situations of the overall water supply network. However, the pressure regulation strategies for macro-level pipe networks cannot be fully applied to the case of secondary water supply. In other words, existing technologies still lack pressure control solutions specifically for secondary water supply. Summary of the Invention

[0006] Based on the aforementioned technical problems in the existing technology, a technical solution for a time-sharing pressure control system and control method is provided, aiming to provide a pressure control solution suitable for secondary water supply systems and solve the pressure regulation problem in secondary water supply systems.

[0007] The above technical solutions specifically include:

[0008] A time-segmented pressure control system is applied to a secondary water supply system; wherein, it includes:

[0009] The data acquisition unit is used to acquire the original historical data of the secondary water supply system and the original pump-related data of each water pump;

[0010] A water volume prediction unit, connected to the data acquisition unit, is used to construct a water volume prediction model based on the original historical data, and to predict water volume prediction data based on the water volume prediction model.

[0011] The pressure prediction unit, connected to the data acquisition unit, is used to construct a pressure prediction model based on the original historical data and to predict import pressure prediction data based on the pressure prediction model.

[0012] A pressure fitting unit, connected to the data acquisition unit, is used to fit a first correspondence between the outlet flow rate and the outlet pressure based on the original historical data.

[0013] A pump fitting unit, connected to the data acquisition unit, is used to fit a second correspondence between the outlet flow rate and the actual operating parameters of the pump based on the original historical data and the pump-related original data.

[0014] The optimization configuration unit is connected to the data acquisition unit, the water volume prediction unit, the pressure prediction unit, the pressure fitting unit, and the water pump fitting unit, respectively. It is used to divide the operation time of the secondary water supply system into time periods based on the daily flow trend data predicted from the original historical data, the pump-related original data, the water consumption prediction data, the outlet pressure prediction data, the first correspondence relationship, and the second correspondence relationship, and to configure corresponding water pump control schemes for each segmented operation period.

[0015] Preferably, in this time-segmented pressure control system, the water volume prediction unit includes:

[0016] The first acquisition module is used to acquire export flow data and various influencing factor data from the original historical data;

[0017] The first model training module, connected to the first acquisition module, is used to perform correlation analysis on the outlet flow data and the influencing factor data. Based on the analysis results, the first N outlet flow data and the multiple influencing factor data with correlation are used as the input data of the training sample, and the outlet flow data as the prediction result are used as the output data of the training sample to train the water volume prediction model.

[0018] The water volume prediction module is connected to the first model training module and is used to input multiple outlet flow data and multiple related influencing factor data, which serve as the basis for prediction, into the trained water volume prediction module to obtain the water volume prediction data.

[0019] Preferably, in this time-segmented pressure control system, the water volume prediction unit further includes:

[0020] The first lag analysis module is connected to the first acquisition module and the first model training module, respectively, and is used to analyze the lag data of the export flow data to the influencing factor data, and add the lag data to the input data of the training sample.

[0021] Preferably, in this time-segmented pressure control system, the pressure prediction unit includes:

[0022] The second acquisition module is used to acquire import pressure data, import and export flow data, and various influencing factor data from the original historical data;

[0023] The second model training module, connected to the second acquisition module, is used to perform correlation analysis on the inlet pressure data, the inlet and outlet flow data, and the influencing factor data respectively. Based on the analysis results, the first N inlet pressure data, the multiple influencing factors with correlation, and the inlet and outlet flow data used as the prediction basis are used as the input data of the training sample, and the inlet pressure data used as the prediction result are used as the output data of the training sample to train the pressure prediction model.

[0024] The pressure prediction module, connected to the second model training module, is used to input multiple inlet pressure data, as well as multiple related influencing factor data and the inlet and outlet flow rates, which serve as the basis for prediction, into the trained pressure prediction module to obtain the inlet pressure prediction data.

[0025] Preferably, in this time-segmented pressure control system, the pressure prediction unit further includes:

[0026] The second lag analysis module is connected to the second acquisition module and the second model training module, respectively. It is used to analyze the lag data of the import pressure data on the influencing factor data and the import and export flow data, and add the lag data to the input data of the training sample.

[0027] Preferably, in this time-segmented pressure control system, the pressure fitting unit includes:

[0028] The third acquisition module is used to acquire inlet and outlet flow data, outlet pressure data, pump set pressure data at the most unfavorable point, pressure setting data at the most unfavorable point, and simulated pipe loss data from the original historical data.

[0029] The pressure reduction range processing module, connected to the third acquisition module, is used to process the corresponding pressure reduction range of the most unfavorable point pressure based on the most unfavorable point pressure data, the most unfavorable point pressure setting data, and the simulated pipe loss data for different outlet flow data.

[0030] The pressure correction module, connected to the pressure reduction range processing module, is used to correct the corresponding outlet pressure data according to the pressure reduction range to obtain reasonable outlet pressure data.

[0031] The relationship construction module, connected to the pressure correction module, is used to fit the first correspondence between the different numerical ranges of the outlet flow data and the numerical range of the reasonable outlet pressure data.

[0032] Preferably, in this time-segmented pressure control system, the water pump fitting unit includes:

[0033] The fourth acquisition module is used to acquire outlet flow rate data, inlet and outlet pressure data, power consumption data and pump operating frequency data from the original historical data, and to acquire pump factory parameter data from the pump-related original data.

[0034] The first processing module, connected to the fourth acquisition module, is used to process the original historical data and the pump-related original data to obtain multiple standard relationship change fitting curves of the water pump under the power frequency.

[0035] The second processing module is connected to the first processing module and the fourth acquisition module respectively. Based on the standard relationship change fitting curve, the original historical data and the pump-related original data, it processes the relationship change fitting curve between the flow rate, head, power and efficiency of the water pump at different operating frequencies, and uses it as the second correspondence.

[0036] Preferably, in this time-segmented pressure control system, the standard relationship variation fitting curve includes:

[0037] Fitted curve of the first standard relationship between flow rate and head;

[0038] The fitted curve of the second standard relationship between flow rate and power; and

[0039] The fitting curve for the change in the third standard relationship between flow rate and efficiency.

[0040] Preferably, in this time-segmented pressure control system, the optimization configuration unit uses a two-layer genetic algorithm to segment the operating time of the secondary water supply system into time segments, and configures corresponding pump control schemes for each segmented operating time segment, wherein:

[0041] The genetic algorithm in the first layer is used to separate the operating segments of the secondary water supply system into different time periods to obtain each operating segment;

[0042] The genetic algorithm in the second layer is used to process and obtain the water pump control scheme with the lowest overall power consumption for each of the running segments, so as to be the water pump control scheme configured within the running segment.

[0043] A time-segmented pressure control method is applied to a secondary water supply system; wherein the method includes:

[0044] Step S1: Obtain the original historical data of the secondary water supply system and the original data related to each water pump.

[0045] Step S2: Construct a water volume prediction model based on the original historical data, predict water volume prediction data based on the water volume prediction model, construct a pressure prediction model based on the original historical data, predict inlet pressure prediction data based on the pressure prediction model, fit a first correspondence between outlet flow rate and outlet pressure based on the original historical data, and fit a second correspondence between outlet flow rate and actual operating parameters of the pump based on the original historical data and the pump-related original data.

[0046] Step S3: Based on the daily flow trend data predicted from the original historical data, the original pump-related data, the predicted water consumption data, the predicted outlet pressure data, the first correspondence, and the second correspondence, the operating time of the secondary water supply system is divided into time periods, and a corresponding water pump control scheme is configured for each segmented operating period.

[0047] The beneficial effects of the above technical solution are as follows: it proposes a more suitable solution for secondary water supply, replacing the constant pressure water supply method in existing technologies, making the time-sharing pressure control of secondary water supply more reasonable and precise. At the same time, by providing a relatively superior time-sharing pump allocation scheme, it can meet users' "high-quality water use" needs while maximizing the operation of water pumps in their high-efficiency range, further reducing energy consumption. Attached Figure Description

[0048] Figure 1 This is a schematic diagram of the overall structure of a time-sharing pressure control system in a preferred embodiment of the present invention.

[0049] Figure 2 This is a schematic diagram of the overall implementation principle of the time-sharing pressure control system in a preferred embodiment of the present invention.

[0050] Figure 3 This is a schematic diagram of the specific structure of the water volume prediction unit in a preferred embodiment of the present invention.

[0051] Figure 4 This is a schematic diagram illustrating the implementation principle of the water volume prediction unit in a preferred embodiment of the present invention.

[0052] Figure 5 This is a schematic diagram of the specific structure of the pressure prediction unit in a preferred embodiment of the present invention;

[0053] Figure 6 This is a schematic diagram illustrating the implementation principle of the pressure prediction unit in a preferred embodiment of the present invention.

[0054] Figure 7 This is a schematic diagram of the specific structure of the pressure fitting unit in a preferred embodiment of the present invention;

[0055] Figure 8 This is a schematic diagram illustrating the implementation principle of the pressure fitting unit in a preferred embodiment of the present invention.

[0056] Figure 9 This is a schematic diagram of the specific structure of the water pump fitting unit in a preferred embodiment of the present invention;

[0057] Figure 10-11 This is a schematic diagram illustrating the implementation principle of the two-layer genetic algorithm set in the optimization configuration unit in a preferred embodiment of the present invention.

[0058] Figure 12 This is a schematic diagram of the overall process of a time-segmented pressure control method in a preferred embodiment of the present invention.

[0059] Figure 13 This is a flowchart illustrating the water volume prediction process in a preferred embodiment of the present invention.

[0060] Figure 14 This is a flowchart illustrating the pressure prediction process in a preferred embodiment of the present invention.

[0061] Figure 15 This is a flowchart illustrating the pressure fitting process in a preferred embodiment of the present invention.

[0062] Figure 16 This is a flowchart illustrating the water pump fitting process in a preferred embodiment of the present invention. Detailed Implementation

[0063] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0064] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0065] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, but this is not intended to limit the scope of the invention.

[0066] In existing technologies, pressure regulation strategies for secondary water supply systems can employ time-segmented pressure control, which typically employs two methods:

[0067] The first method is to intuitively set fixed time periods based on existing data, such as peak and off-peak periods, or to divide the operating time into more detailed time periods, such as peak, secondary peak, trough, secondary trough, stable operating period, and secondary stable operating period. Then, constant pressure water supply is set for each of these segmented time periods to achieve the purpose of time-segmented pressure control.

[0068] The second approach involves dividing the existing data according to a certain algorithm, conducting comprehensive analysis using multiple pressure points, dividing the data into multiple fixed time periods, and then using various pressure regulation strategies for pressure regulation control based on the scenario.

[0069] The first method described above uses fixed time periods for different scenarios, but the actual control periods may vary each day. This can lead to a situation where the fixed time periods do not match the actual operation of the secondary water supply system, resulting in the pressure regulation strategy not being able to fully adapt to the actual operation of the secondary water supply system.

[0070] The second method described above, while employing comprehensive analysis of multiple pressure points and capable of dividing the secondary water supply system into different fixed time periods to some extent, relies excessively on actual data. This means that while the final fixed time periods may match the actual data, they may not reflect the daily flow trend of the secondary water supply system. Furthermore, dividing the system into too many fixed time periods leads to frequent pump switching, accelerating pump and related equipment wear and tear; conversely, dividing it into too few fixed time periods results in excessively high overall energy consumption.

[0071] Therefore, based on the aforementioned technical problems mentioned in the prior art, a time-sharing pressure control system is now provided. This time-sharing pressure control system is applied to a secondary water supply system, and its specific details are as follows: Figure 1 As shown, it includes:

[0072] Data acquisition unit 1 is used to acquire the original historical data of the secondary water supply system and the pump-related original data of each water pump;

[0073] Water volume prediction unit 2 is connected to data acquisition unit 1 and is used to build a water volume prediction model based on the original historical data and to predict water volume prediction data based on the water volume prediction model.

[0074] Pressure prediction unit 3 is connected to data acquisition unit 1 and is used to build a pressure prediction model based on the original historical data and to predict import pressure prediction data based on the pressure prediction model.

[0075] Pressure fitting unit 4 is connected to data acquisition unit 1 and is used to fit the first correspondence between outlet flow rate and outlet pressure based on the original historical data.

[0076] The water pump fitting unit 5 is connected to the data acquisition unit 1 and is used to fit the second correspondence between the outlet flow rate and the outlet pressure based on the original historical data and the pump-related original data.

[0077] The optimization configuration unit 6 is connected to the data acquisition unit 1, water volume prediction unit 2, pressure prediction unit 3, pressure fitting unit 4 and water pump fitting unit 5 respectively. It is used to divide the operation time of the secondary water supply system into time periods based on the daily flow trend data predicted from the original historical data, pump-related original data, water volume prediction data, outlet pressure prediction data, first correspondence relationship and second correspondence relationship, and configure the corresponding water pump control scheme for each segmented operation period.

[0078] Specifically, in this embodiment, dividing the pump's operating period based on historical data avoids the problem of dividing the fixed time period too much or too little. Furthermore, estimating the daily pressure trend based on a water volume prediction algorithm avoids over-reliance on real-time data. Additionally, predicting the water demand trend and pressure range based on water volume prediction, pressure prediction, pressure fitting, and pump fitting allows for guidance on future pump operation using a pump allocation model.

[0079] In this embodiment, the working principle of the above-mentioned time-sharing pressure control system is as follows: Figure 2 As shown, data acquisition unit 1 acquires raw historical data and pump-related raw data, wherein:

[0080] Water volume prediction unit 2 constructs a water volume prediction model based on the original historical data, and predicts the water consumption data for the day based on the water volume prediction model.

[0081] Pressure prediction unit 3 constructs a pressure prediction model based on the original historical data, and predicts the import pressure prediction data for the day based on the pressure prediction model.

[0082] Pressure fitting unit 4 obtains the first correspondence between outlet flow rate and outlet pressure based on the original historical data. Subsequently, based on this first correspondence, the reasonable range of outlet pressure values ​​corresponding to the daily flow rate data range can be determined.

[0083] The pump fitting unit 5 fits the original historical data and pump-related original data to obtain a second correspondence between the outlet pressure and the outlet water flow. This second correspondence is actually a flow rate variation curve for head, power and efficiency, respectively. Subsequently, the pump head, power and efficiency can be determined through this variation curve and the actual flow rate data.

[0084] The optimization configuration unit 6 can divide the daily operating period into time segments based on the above water consumption prediction data, inlet pressure prediction data, first correspondence, second correspondence, daily flow trend data predicted through original historical data, and pump-related original data. It can then configure a control scheme for each water pump for each time segment and finally control the operation of each water pump in each time segment according to the time segment control scheme.

[0085] In this embodiment, the above-mentioned daily traffic trend data can be predicted from the original historical data based on the existing traffic prediction model, which will not be elaborated here.

[0086] In a preferred embodiment of the present invention, the aforementioned original historical data may include: inlet flow rate data, outlet flow rate data, inlet pressure data, outlet pressure data, the most unfavorable pressure data of the pump set, pump operating frequency data, and other key data related to the pump set. Wherein:

[0087] Since the data of individual pump sets in a secondary water supply system are basically the same, the inlet flow rate data, outlet flow rate data, inlet pressure data, and outlet pressure data are all the same within a single pump set, and will not be distinguished in the following text.

[0088] The so-called pressure data of the most unfavorable point of the pump set refers to the pressure data at the farthest point of the water supply of the pump set. Specifically, it can include the reading data of the pressure gauge installed at the highest point of the water supply pipeline of the residential building at the farthest point of the water supply of the pump set.

[0089] The so-called other key data related to the pump set can include pump set-related configuration data and simulation data. The pump set-related configuration data includes the setting data for the pump set's most unfavorable pressure point. The pump set-related simulation data includes pipe loss data simulated based on historical real data and pipe loss data simulated based on theoretical data. Pipe loss data refers to the water supply data (flow rate, pressure, etc.) that can be achieved at a certain water supply point but cannot be achieved due to pipe loss in the pipeline where the water supply point is located. This type of pipe loss data can be simulated using historical real data or theoretical data and used in subsequent calculations.

[0090] It is worth noting that the simulated pipe loss data obtained above will differ under different flow rates and pressure conditions. Specifically, this simulated pipe loss data can be obtained by fitting the data using the following formula: H 水泵扬程 =(H 高差 +H 出水头 +H 管损 )*1.1-H 上端来水压力 By using the determined elevation difference, the known upstream water pressure, the pump head, and the pressure at the most unfavorable point (outlet head), pipe loss data under different flow rates can be fitted, thus fitting the correspondence between flow rate and simulated pipe loss data.

[0091] In a preferred embodiment of the present invention, the aforementioned original historical data further includes various influencing factor data. Influencing factor data refers to external factors that can affect the secondary water supply system, such as weather data, holiday data, and special event data. Wherein:

[0092] Weather data includes temperature, humidity, and precipitation data.

[0093] The holiday factor data includes data for determining whether it is a workday, a rest day, or a major holiday.

[0094] Special event factor data can include external factors that cause a sudden increase in water consumption and external factors that cause a sudden decrease in water consumption. For example, holding a large-scale event temporarily may cause a sudden increase in water consumption in a certain place, while a large number of people returning to their hometowns from a first-tier city during the Spring Festival may cause a rapid decrease in water consumption in that first-tier city.

[0095] In a preferred embodiment of the present invention, the aforementioned pump-related raw data refers to data related to the pump itself, such as the pump model, factory parameters, etc. The factory parameters may include data such as the pump's rated power and rated head, which will not be elaborated here.

[0096] It is worth noting that, for the sake of consistency throughout the calculation process, all the aforementioned raw historical data and pump-related historical data need to be represented by corresponding numerical values. In particular, it is important to note that for some status-related data, such as influencing factor data, different numerical values ​​can be used to label different statuses. For example, 0 can be used to label a workday, 1 to label a rest day, and 2 to label a major holiday. These details will not be elaborated further here.

[0097] In a preferred embodiment of the present invention, such as Figure 3 As shown, the water volume prediction unit 2 specifically includes:

[0098] The first acquisition module 21 is used to acquire export flow data and various influencing factor data from the original historical data;

[0099] The first model training module 22 is connected to the first acquisition module 21 and is used to perform correlation analysis on the outlet flow data and the influencing factor data. Based on the analysis results, the first N outlet flow data and the data of multiple influencing factors with correlation are used as the input data of the training sample, and the outlet flow data as the prediction result is used as the output data of the training sample to train the water volume prediction model.

[0100] The water volume prediction module 23 is connected to the first model training module 22. It is used to input multiple outlet flow data and multiple related influencing factor data, which serve as the basis for prediction, into the trained water volume prediction module to obtain water volume prediction data.

[0101] Specifically, in this embodiment, for the water volume prediction process, it is first necessary to train a water volume prediction model. The specific method for constructing the training sample data of the water volume prediction model is as follows:

[0102] The first acquisition module 21 acquires export flow data and various influencing factor data from the original historical data.

[0103] After obtaining the aforementioned raw historical data, data preprocessing is performed to filter out valid raw historical data. The purpose of data preprocessing is to supplement missing data, eliminate extreme data and noisy data, that is, to preprocess the outlier data in the historical data.

[0104] For water consumption data in the original historical data (original historical data that is not considered influencing factor data, such as inlet and outlet flow data and inlet and outlet pressure data), the data preprocessing process can be carried out in the following way:

[0105] First, obtain the abnormal data time points in the original historical data. At these abnormal data time points, there may be missing or abnormal data, meaning that there is no usable and valid original historical data at these abnormal data time points.

[0106] Next, it is determined whether original predicted data exists at the time point of the anomaly. Original predicted data refers to the data at the time point of the anomaly that can be predicted based on historical data preceding the anomaly; this original predicted data is also recorded in the historical data. Since original predicted data is actually a part of a certain type of historical data—in other words, a certain type of historical data can actually include both the actual and predicted data values ​​of that historical data—this will not be elaborated further below. Therefore:

[0107] If there is corresponding original predicted data at the time point of the abnormal data, and the original predicted data is valid, then the original predicted data will be used to fill in the data at the time point of the abnormal data.

[0108] If there is no corresponding original prediction data at the time point of the abnormal data, that is, there is no corresponding original prediction data at the time point of the abnormal data, or although there is original prediction data at the time point of the abnormal data, the original prediction data is still abnormal data (unavailable), then the mean of at least 7 data points under other influencing factor data that are similar to the influencing factor data corresponding to the time point of the abnormal data will be used to fill the gap.

[0109] For data on various influencing factors in the original historical data, data preprocessing can be performed in the following ways:

[0110] Abnormal data in influencing factor data mainly includes the following three types:

[0111] Weather factor data missing: Weather factor data values ​​are empty or the data does not exist;

[0112] Weather data is too extreme: Weather data values ​​exceed the normal range;

[0113] Holiday / Special Event Data Missing: Holiday / Special Event Data Values ​​are empty or do not exist.

[0114] Furthermore:

[0115] When outlier data is due to missing weather factors or excessively extreme weather conditions, the following methods can be used to handle it:

[0116] 1. If the amount of weather factor data is less than the amount of export flow data (i.e., the data is not only missing but also mismatched), then the abnormal weather factor data will not be processed and will not be included in the scope of influencing factors.

[0117] 2. If the amount of weather factor data is not less than the amount of outflow data (i.e., the data only has anomaly and missing data), then the average of the two data points before and after the anomalous weather factor data point will be used to fill the gap, replacing the anomalous weather factor data point and participating in the calculation.

[0118] When the abnormal data is due to holiday factors or special event factors, it should be filled with relevant data from normal working days or without special events.

[0119] After the above data preprocessing, effective raw historical data can be obtained and training samples for subsequent water volume prediction models can be constructed.

[0120] In this invention, in the historical prediction unit 3, pressure fitting unit 4, and water pump fitting unit 5, all those involving the use of original historical data need to undergo data preprocessing on the acquired original historical data to obtain valid original historical data. The data preprocessing method can be referred to above, and will not be described in detail in this invention.

[0121] In this embodiment, after obtaining valid original historical data, the export flow data and various influencing factor data in the original historical data are first subjected to correlation analysis. The correlation analysis adopts Pearson correlation analysis, which includes the correlation coefficient r and the significance level p. The formula for calculating the correlation coefficient r is as follows:

[0122]

[0123] in,

[0124] i = 1, 2, 3, ..., n, which represents the sequence number of the original historical data at multiple different times involved in the calculation.

[0125] X i X is used to represent the i-th valid influencing factor data, and X is used to represent the average value of all influencing factor data involved in the calculation.

[0126] Y i Y is used to represent the i-th valid export flow data, and Y is used to represent the average value of all export flow data involved in the calculation.

[0127] The significance level p is calculated as follows: Assuming H0 is R=0, there is no linear relationship between the two variables (export flow data and influencing factor data), then the significance level p is calculated.

[0128] The range of the correlation coefficient r above is divided as follows:

[0129] If 0.8 < r ≤ 1.0, it is considered to be extremely strong correlation;

[0130] If 0.6 < r ≤ 0.6, a strong correlation is considered.

[0131] If 0.2 < r ≤ 0.4, it is considered a weak correlation;

[0132] If 0 ≤ r ≤ 0.2, it is considered to be a very weak correlation or no correlation.

[0133] The range of results for the above significance level p is divided as follows:

[0134] p < 0.05 indicates that the two data columns are significantly correlated.

[0135] If p ≥ 0.05, the two columns of data are considered to be uncorrelated.

[0136] In this embodiment, the prediction principle of the above-mentioned water volume prediction model is as follows: The current outlet flow rate is predicted based on the top N outlet flow rate data and the influencing factor data related to the corresponding outlet flow rate data. For example, if it is necessary to combine the outlet flow rate data of the previous 30 days and their influencing factor data to predict the outlet flow rate data of the current day, then in the training sample of the corresponding water volume prediction model, the output data is the outlet flow rate data of a certain day in the original historical data (i.e., the outlet flow rate data as the prediction result mentioned above), and the input data is all the outlet flow rate data of the previous 30 days and their related influencing factor data. Further, the influencing factor data associated with the daily outlet flow rate data in the input data can be determined according to the above-mentioned correlation analysis method. Specifically, the correlation analysis results are obtained by combining the above-mentioned correlation coefficient r and significance level p. Based on the correlation analysis results, multiple influencing factor data with the highest ranking are selected and added to the exogenous variables of the daily outlet flow rate data as input data for the training sample. For example, multiple influencing factor data can be selected from top to bottom and added to the exogenous variables associated with the daily outlet flow rate data, or the influencing factor data can be selected based on the numerical range of the correlation coefficient and significance level. For example, for the correlation coefficient r, select "extremely strong correlation" and "strong correlation"; for the significance level p, select two columns of data that are significantly correlated; select the influencing factor data that simultaneously meets the above two conditions and add them to the exogenous variables of the daily export flow data as input data for the training sample.

[0137] It is worth noting that when conducting correlation analysis on influencing factors, the export flow should be analyzed separately with the data of each type of influencing factor, rather than combining the data of influencing factors together for analysis.

[0138] Furthermore, some influencing factor data do not immediately cause changes in export flow data; that is, there is a certain lag between the influencing factor data and the export flow data, which may be minutes or even hours. Therefore, in order to more accurately reflect the correlation between influencing factor data and export flow data, in addition to correlation analysis, lag analysis is also needed. The corresponding lag data is added to the exogenous variables of the daily export flow data as input data for training samples. Therefore, in the preferred embodiment of the present invention, as... Figure 3 As shown, the water volume prediction unit 2 also includes:

[0139] The first lag analysis module 24 is connected to the first acquisition module 21 and the first model training module 22 respectively. It is used to analyze the lag data of the export flow data on the influencing factor data and add the lag data to the input data of the training sample.

[0140] Specifically, lag analysis can be performed using the Granger causality test, the formula of which is expressed as follows:

[0141]

[0142]

[0143] in,

[0144] x is used to represent influencing factor data, x t Data used to represent influencing factors in the current raw historical data;

[0145] y is used to represent export flow data, y t Used to represent export flow data in the current raw historical data;

[0146] In formula (2), i = 1, 2, 3, ..., q;

[0147] In formula (3), i = 1, 2, 3, ..., s;

[0148] α, λ, β, and δ are all parameters;

[0149] u 1t and u 2t Both are used to represent white noise, and they are unrelated.

[0150] In this embodiment, based on the above formulas (2) and (3), the lag data of the export flow data relative to the influencing factor data can be calculated, and this lag data is also added to the input data of the training sample, that is, added to the exogenous variables. Of course, the lag analysis is also performed separately for each influencing factor data.

[0151] Finally, based on the above implementation principle, training sample data for the water volume prediction model is constructed. The input data for the training samples consists of the first N outlet flow rates as the basis for prediction, along with the constructed exogenous variables (influencing factor data and lag data) associated with each outlet flow rate. The output data is the outlet flow rate data for a specific day, which serves as the prediction result. The model is then trained using the constructed training samples to obtain the finally trained water volume prediction model.

[0152] In practical use, the outflow data of the previous N days, which serve as the basis for prediction, as well as the data of multiple influencing factors that are related to the daily outflow data, are input into the water volume prediction model to predict the outflow data of the day, which is to say, the water consumption prediction data.

[0153] The specific implementation principle of the above-mentioned water volume prediction unit 2 can be referred to Figure 4 .

[0154] In a preferred embodiment of the present invention, such as Figure 5 As shown, the pressure prediction unit 3 includes:

[0155] The second acquisition module 31 is used to acquire import pressure data, import and export flow data, and various influencing factor data from the original historical data;

[0156] The second model training module 32 is connected to the second acquisition module 31. It is used to perform correlation analysis on the import pressure data, import and export flow data and influencing factor data respectively. Based on the analysis results, the first N import pressure data, multiple influencing factors with correlation and import and export flow data used as the prediction basis are used as the input data of the training sample, and the import pressure data used as the prediction result is used as the output data of the training sample to train the pressure prediction model.

[0157] The pressure prediction module 33 is connected to the second model training module 32. It is used to input multiple inlet pressure data, as well as multiple related influencing factor data and inlet and outlet flow rates, which serve as the basis for prediction, into the trained pressure prediction module to obtain inlet pressure prediction data.

[0158] Specifically, the implementation in this embodiment is similar to the water volume prediction unit described above. The difference lies in the fact that the exogenous variables of the daily inlet pressure data in the training sample include not only various related influencing factor data but also related inlet and outlet flow data (inlet flow data and outlet flow data). The output data of the training sample is the inlet pressure data for a specific day to be predicted. Therefore, when performing correlation analysis, the implementation method in the water volume prediction unit can be analogous, i.e., Pearson correlation analysis is used to analyze the correlation between each influencing factor data and the inlet pressure data, as well as the correlation between the inlet and outlet flow data and the inlet pressure data, to obtain the correlation coefficient r and significance level p. Subsequently, based on the results of the correlation analysis, several influencing factor data and inlet and outlet flow data with the highest correlation are selected to be included in the exogenous variables of the daily inlet pressure data in the input data of the training sample, thereby constructing the input data of the training sample.

[0159] In this embodiment, it is still as follows Figure 5 As shown, the pressure prediction unit 3 also includes:

[0160] The second lag analysis module 34 is connected to the second acquisition module 31 and the second model training module 32, respectively. It is used to analyze the lag data of the import pressure data on the influencing factor data and the import and export flow data, and add the lag data to the input data of the training sample. This lag analysis method can be analogous to the above formulas (2) and (3), and is obtained by Granger causality detection. The difference is that in this embodiment, the exogenous variables of the daily import pressure data include not only the influencing factor data but also the import and export flow data. Therefore, when performing lag analysis, in addition to analyzing each influencing factor data separately, it is also necessary to analyze the import and export flow data individually to obtain the lag data and include it in the exogenous variables of the daily import pressure data in the input data of the training sample.

[0161] Finally, based on the above implementation principle, training sample data for the pressure prediction model is constructed. The input data for the training samples consists of the first N import pressure data points used as the basis for prediction, along with the constructed exogenous variables associated with each import pressure data point (influencing factor data + lagged data, and import / export flow data + lagged data). The output data is the import pressure data for a specific day, serving as the prediction result. The model is then trained using the constructed training samples to obtain the finally trained pressure prediction model.

[0162] In practical use, the import pressure data of the previous N days, the data of multiple influencing factors that are related to the daily import pressure data, and the import and export flow data are input into the pressure prediction model to predict the import pressure data of the day, that is, to obtain the import pressure prediction data.

[0163] The specific implementation principle of the pressure prediction unit 3 mentioned above can be referred to Figure 6 .

[0164] In a preferred embodiment of the present invention, such as Figure 7 As shown, the pressure fitting unit 4 includes:

[0165] The third acquisition module 41 is used to acquire inlet and outlet flow data, outlet pressure data, pump set related most unfavorable point pressure data, most unfavorable point pressure setting data and simulated pipe loss data from the original historical data.

[0166] The pressure reduction range processing module 42 is connected to the third acquisition module 41. It is used to process the corresponding pressure reduction range of the most unfavorable point pressure based on the most unfavorable point pressure data, the most unfavorable point pressure setting data and the simulated pipe loss data for different outlet flow data.

[0167] The pressure correction module 43 is connected to the pressure reduction range processing module 42 and is used to correct the corresponding outlet pressure data according to the pressure reduction range in order to obtain reasonable outlet pressure data.

[0168] The relationship building module 44 is connected to the pressure correction module 43, and is used to fit the first correspondence between the numerical range of different outlet flow data and the numerical range of reasonable outlet pressure data.

[0169] Specifically, in this embodiment, as mentioned above, the so-called most unfavorable point refers to the residential building at the farthest end of the water supply from the pump set, and the reading of the pressure gauge at the highest point of the water supply from the pump set. The pressure data at the most unfavorable point refers to the actual reading of the pressure gauge, and the pressure setting data at the most unfavorable point refers to the theoretical value of the pressure at the pre-set most unfavorable point.

[0170] In this embodiment, the inlet and outlet flow data, outlet pressure data, and pressure data at the most unfavorable point of the pump set are obtained from the original historical data. After preprocessing these data, valid data is obtained and used in subsequent calculations.

[0171] Subsequently, simulated pipe loss data (calculated as described above) and the setting data of the most unfavorable point pressure are obtained from the original historical data. Since these data are obtained from simulation calculations or pre-set, they do not need to be pre-processed.

[0172] In this embodiment, based on different inlet and outlet flow rate ranges, the pressure reduction range at the most unfavorable point is calculated for each range. The specific calculation method for this pressure reduction range is as follows:

[0173] Within the numerical range of each outlet flow data, there should be a corresponding numerical range of the most unfavorable point pressure. By subtracting the preset most unfavorable point pressure setting data from the highest and lowest values ​​of this most unfavorable point pressure range, a pressure reduction range for the most unfavorable point pressure can be obtained.

[0174] Subsequently, based on the pressure reduction range of the most unfavorable point, combined with simulated pipe loss data under different flow rates and pressures, and outlet pressure data from the original historical data, a reasonable range of pump set outlet pressure can be derived. Therefore, based on the range of each outlet flow rate, a reasonable range of pump set outlet pressure can be obtained, thus establishing the first correspondence between outlet flow rate and pump set outlet pressure. Then, based on the actual daily outlet flow rate data, the reasonable outlet pressure data for that flow rate can be obtained.

[0175] The specific implementation principle of the pressure fitting unit 4 mentioned above can be referred to Figure 8 .

[0176] In a preferred embodiment of the present invention, such as Figure 9 As shown, the above-mentioned water pump fitting unit 5 includes:

[0177] The fourth acquisition module 51 is used to acquire outlet flow rate data, inlet and outlet pressure data, power consumption data and pump operating frequency data from the original historical data, as well as pump factory parameter data from the pump-related original data.

[0178] The first processing module 52 is connected to the fourth acquisition module 51 and is used to process the original historical data and pump-related original data to obtain the fitting curves of multiple standard relationship changes of the water pump under the power frequency.

[0179] The second processing module 53 is connected to the first processing module 52 and the fourth acquisition module 51 respectively. Based on the standard relationship change fitting curve, the original historical data and the pump-related original data, it processes the relationship change fitting curve between the flow rate, head, power and efficiency of the water pump at different operating frequencies, and uses it as the second correspondence.

[0180] Specifically, in this embodiment, outlet flow rate data, inlet and outlet pressure data, power consumption data, and pump operating frequency data are first obtained from the original historical data, and pump factory parameter data are obtained from the pump-related raw data. The aforementioned original historical data needs to undergo data preprocessing to obtain valid data for subsequent calculations, while the aforementioned pump-related raw data does not require data preprocessing.

[0181] In this embodiment, the above-mentioned multiple standard relationship change fitting curves respectively include:

[0182] Fitted curve of the first standard relationship between flow rate and head;

[0183] The fitted curve of the second standard relationship between flow rate and power; and

[0184] The fitting curve for the change in the third standard relationship between flow rate and efficiency.

[0185] Specifically, by first calculating the original historical data and pump-related raw data obtained above, we can obtain the inlet and outlet flow rates, head, efficiency, and power of the water pump at certain discrete operating frequencies. These data can be converted into standard relationship fitting curves for flow rate-head, flow rate-efficiency, and flow rate-power at the power frequency.

[0186] The fitting curves for the changes in the above three standard relationships were obtained by converting them using the pump proportionality law, namely, the flow proportionality law Q1 / Q2=n1 / n2 and the head proportionality law H1 / H2=(n1 / n2). 3 And the shaft power proportionality law P1 / P2=(n1 / n2) 3The result is obtained through conversion. Among them:

[0187] The fitting curve for the first standard relationship between flow rate and head can be expressed in the following form:

[0188] H = A2Q 2 +A1Q+A0; (4)

[0189] Where H represents the head, Q represents the inlet and outlet flow rates, and A0, A1, and A2 are parameter values ​​obtained by fitting the above-mentioned original historical data and pump-related original data.

[0190] The fitted curve of the second standard relationship between flow rate and power can be expressed in the following form:

[0191] P = B2Q 2 +B1Q+B0; (5)

[0192] Where P represents power, and B0, B1, and B2 are parameter values, which are obtained by fitting the above-mentioned original historical data and pump-related original data.

[0193] The fitting curve for the third-standard relationship between flow rate and efficiency can be expressed in the following form:

[0194] η=C2Q 2 +C1Q+C0; (6)

[0195] Where η represents efficiency, and C0, C1, and C2 are parameter values, which are obtained by fitting the above original historical data and pump-related original data.

[0196] After fitting the above three standard relationship change fitting curves, we can obtain the corresponding three relationship change fitting curves of the water pump at different operating frequencies, which can be used as the second correspondence.

[0197] Specifically, the fitted curve of the flow rate-head relationship can be obtained by the following conversion method:

[0198] If we assume f represents different operating frequencies. According to the proportional law of water pumps mentioned above, we have:

[0199]

[0200]

[0201] According to formulas (4), (7) and (8), we can obtain:

[0202] H1 = S 2 ×H2=(A2Q2 2 +A1Q2+A0)×S2 (9)

[0203]

[0204] Furthermore, according to the above formulas (9) and (10), we can obtain:

[0205]

[0206] By rearranging the above formula (11), we finally obtain the fitting curve of the flow-head relationship when the operating frequency is f:

[0207]

[0208] Similarly, the flow-power relationship curve (formula (13) below) and the flow-efficiency relationship curve (formula (14) below) can be obtained by converting the above formulas:

[0209]

[0210]

[0211] In this embodiment, the three fitting curves (Formulas 12-14) of the flow-head, flow-power, and flow-efficiency relationships obtained at different operating frequencies are used as the second corresponding relationship output.

[0212] It should be noted that the three relationship change fitting curves obtained at different operating frequencies need to be corrected by the original historical data and pump-related original data to better reflect the actual situation.

[0213] In a preferred embodiment of the present invention, the optimization configuration unit 6 employs a two-layer genetic algorithm to segment the operating time of the secondary water supply system into time periods, and configures corresponding water pump control schemes for each segmented operating period, wherein:

[0214] The first layer of the genetic algorithm is used to separate the operating segments of the secondary water supply system into different time periods to obtain each operating segment;

[0215] The second layer of the genetic algorithm is used to process and obtain the water pump control scheme with the lowest overall power consumption for each running segment, which is then used as the water pump control scheme configured within the running segment.

[0216] Specifically, in this embodiment, the specific implementation principle of the above-mentioned optimized configuration unit 6 is described in [reference needed]. Figure 10 The input parameters include head prediction data, water consumption prediction data, first correspondence, second correspondence, pump-related raw data, and daily flow prediction data, among which:

[0217] Daily flow forecast data can also be obtained using water volume forecasting models. The difference between daily flow forecast data and water consumption forecast data lies in the following: water consumption forecast data provides specific numerical values, representing a static representation of water usage, such as predicting the exact amount of water used on a given day. Daily flow forecast data, on the other hand, can represent a dynamic range of numerical changes, reflecting a dynamic water consumption pattern. For example, it can be used to predict whether water consumption will increase or decrease in the next moment, or to predict peak and off-peak periods of water consumption. These details will not be elaborated upon further here.

[0218] The head prediction data actually refers to the outlet pressure data minus the inlet pressure data. That is, the reasonable pump set outlet pressure is obtained based on the daily flow prediction data and the first correspondence, and then the daily inlet pressure prediction data obtained by the pressure prediction model is subtracted to obtain the daily head prediction data.

[0219] Based on the above input parameters and the pre-set two-layer genetic algorithm, the above-mentioned optimization configuration unit 6 finally outputs multiple running segments obtained by dividing the total running segments of a day, as well as the water pump control scheme configured under each running segment.

[0220] Furthermore, such as Figure 11 As shown, the time-segmentation algorithm used by the optimization configuration unit 6 is specifically a two-layer genetic algorithm, wherein:

[0221] The purpose of the first-layer genetic algorithm is to obtain reasonable time segmentation points and reasonable pump allocation results. The number of genes is the range of inlet and outlet flow rates, and the value of the gene locus indicates whether the current inlet and outlet flow rates are time segmentation points. When constructing the initial population, each flow rate value point is traversed, and each flow rate value point is regarded as a possible time segmentation point. The constraints are that the current outlet pressure data is the maximum outlet pressure in the segmented running period, the fitness is the total power consumption of one day, and the optimal result is the minimum total power consumption of one day.

[0222] Accordingly, the overall daily power consumption can be calculated using a second-layer genetic algorithm. Specifically, the second-layer algorithm calculates the pump control scheme that minimizes power consumption under the given flow and pressure conditions at the current time segment. The number of genes in the second-layer genetic algorithm is twice the number of pumps, representing whether each pump is running and its operating frequency. When constructing the initial population, it is ensured that each pump is likely to run and its operating frequency has a sufficiently wide range. The constraints are that the operating frequency and flow rate of each pump are within their respective settings, and the overall flow data meets the water demand. The flow and pressure data required for the second-layer genetic algorithm calculation are the same as those determined during the time segment calculation using the first-layer genetic algorithm.

[0223] The fitness calculation of the second layer of the genetic algorithm is achieved through the second correspondence output in the pump fitting unit 5 above. That is, after the flow rate, pressure and operating frequency data are known, the corresponding efficiency and power data can be calculated, thereby calculating the corresponding power consumption data, so that the pump can operate in the high-efficiency range as much as possible.

[0224] In summary, based on the two-layer genetic algorithm, it is possible to determine how to divide the overall time period into various running segments when the overall power consumption is minimized, and how to configure the water pump control scheme within each running segment, thereby forming the overall water pump control scheme for the day.

[0225] In a preferred embodiment of the present invention, based on the above-described time-sharing pressure control system, a time-sharing pressure control method is provided, which is applied to a secondary water supply system, and the steps are as follows: Figure 12 As shown, it includes:

[0226] Step S1: Obtain the original historical data of the secondary water supply system and the original data related to each water pump.

[0227] Step S2: Construct a water volume prediction model based on the original historical data, predict water volume prediction data based on the water volume prediction model, construct a pressure prediction model based on the original historical data, predict inlet pressure prediction data based on the pressure prediction model, fit the first correspondence between outlet flow rate and outlet pressure based on the original historical data, and fit the second correspondence between outlet flow rate and actual working parameters of water pump based on the original historical data and pump-related original data.

[0228] Step S3: Based on the daily flow trend data predicted from the original historical data, the original pump data, the water consumption prediction data, the outlet pressure prediction data, the first correspondence relationship, and the second correspondence relationship, the operating time of the secondary water supply system is divided into time periods, and a corresponding water pump control scheme is configured for each segmented operating period.

[0229] In a preferred embodiment of the present invention, such as Figure 13 As shown in the figure, the process of constructing a water volume prediction model based on the original historical data and predicting water volume prediction data based on the water volume prediction model specifically includes:

[0230] Step S21a: Obtain export flow data and various influencing factor data from the original historical data;

[0231] Step S22a: Perform correlation analysis on the outlet flow data and influencing factor data. Based on the analysis results, use the top N outlet flow data and multiple influencing factor data that are related to the prediction as the input data of the training sample, and use the outlet flow data that is the prediction result as the output data of the training sample to train the water volume prediction model.

[0232] Step S23a involves inputting multiple outlet flow data and related influencing factor data, which serve as the basis for prediction, into the trained water volume prediction module to obtain water volume prediction data.

[0233] In a preferred embodiment of the present invention, such as Figure 14 As shown in the diagram, step S2 above, which involves constructing a pressure prediction model based on the original historical data and then using that model to predict import pressure data, specifically includes:

[0234] Step S21b: Obtain import pressure data, import and export flow data, and data on various influencing factors from the original historical data;

[0235] Step S22b: Correlation analysis is performed on the import pressure data, import and export flow data, and influencing factor data respectively. Based on the analysis results, the top N import pressure data, multiple influencing factors with correlation, and import and export flow data used as the prediction basis are used as the input data of the training sample, and the import pressure data used as the prediction result is used as the output data of the training sample to train the pressure prediction model.

[0236] Step S23b involves inputting multiple import pressure data, related influencing factor data, and import / export flow data, which serve as the basis for prediction, into the trained pressure prediction module to obtain import pressure prediction data.

[0237] In a preferred embodiment of the present invention, such as Figure 15 As shown in the figure, the process of fitting the first correspondence between export pressure and import / export flow rate based on the original historical data in step S2 above specifically includes:

[0238] Step S21c: Obtain inlet and outlet flow data, outlet pressure data, pump set pressure data at the most unfavorable point, pressure setting data at the most unfavorable point, and simulated pipe loss data from the original historical data.

[0239] Step S22c: For different inlet and outlet flow data, the pressure reduction range of the corresponding unfavorable point pressure is obtained by processing the most unfavorable point pressure data, the most unfavorable point pressure setting data, and the simulated pipe loss data.

[0240] Step S23c: Correct the corresponding outlet pressure data according to the pressure reduction range to obtain reasonable pump set outlet pressure data;

[0241] Step S24c: Fit the first correspondence between the numerical range of different outlet flow rate data and the numerical range of reasonable pump set outlet pressure data.

[0242] In a preferred embodiment of the present invention, such as Figure 16 As shown in the figure, the process of fitting the second correspondence between the outlet flow rate and the actual operating parameters of the water pump based on the original historical data and the pump-related original data in step S2 specifically includes:

[0243] Step S21d: Obtain the outlet flow rate data, inlet and outlet pressure data, power consumption data, and pump operating frequency data from the original historical data, as well as the pump factory parameter data from the pump-related original data.

[0244] Step S22d: Based on the original historical data and pump-related original data, the fitting curves of multiple standard relationship changes of the water pump under the power frequency are obtained.

[0245] Step S23d: Based on the standard relationship change fitting curve, the original historical data and pump-related original data processing, the relationship change fitting curve between the flow rate, head, power and efficiency of the water pump at different operating frequencies is obtained, which serves as the second correspondence.

[0246] The above description is merely a preferred embodiment of the present invention and does not limit the implementation and protection scope of the present invention. Those skilled in the art should realize that any equivalent substitutions and obvious changes made based on the description and illustrations of the present invention should be included within the protection scope of the present invention.

Claims

1. A time-segmented pressure control system, applied to a secondary water supply system; characterized in that, include: The data acquisition unit is used to acquire the original historical data of the secondary water supply system and the original pump-related data of each water pump; A water volume prediction unit, connected to the data acquisition unit, is used to construct a water volume prediction model based on the original historical data, and to predict water volume prediction data based on the water volume prediction model. The pressure prediction unit, connected to the data acquisition unit, is used to construct a pressure prediction model based on the original historical data and to predict import pressure prediction data based on the pressure prediction model. A pressure fitting unit, connected to the data acquisition unit, is used to fit a first correspondence between the outlet flow rate and the outlet pressure based on the original historical data. A pump fitting unit, connected to the data acquisition unit, is used to fit a second correspondence between the outlet flow rate and the actual operating parameters of the pump based on the original historical data and the pump-related original data. The optimization configuration unit is connected to the data acquisition unit, the water volume prediction unit, the pressure prediction unit, the pressure fitting unit, and the water pump fitting unit, respectively. It is used to divide the operation time of the secondary water supply system into time periods based on the daily flow prediction data obtained from the original historical data, the pump-related original data, the water consumption prediction data, the inlet pressure prediction data, the first correspondence relationship, and the second correspondence relationship, and to configure corresponding water pump control schemes for each segmented operation period. The pressure fitting unit includes: The third acquisition module is used to acquire the outlet flow rate data, outlet pressure data, pump station related most unfavorable point pressure data, most unfavorable point pressure setting data, and simulated pipe loss data from the original historical data. The pressure reduction range processing module, connected to the third acquisition module, is used to process the corresponding pressure reduction range of the most unfavorable point pressure based on the most unfavorable point pressure data, the most unfavorable point pressure setting data, and the simulated pipe loss data for different outlet flow data. The pressure correction module, connected to the pressure reduction range processing module, is used to correct the corresponding outlet pressure data according to the pressure reduction range to obtain reasonable outlet pressure data. A relationship building module, connected to the pressure correction module, is used to fit the first correspondence between the numerical ranges of different outlet flow data and the numerical ranges of the reasonable outlet pressure data. The optimization configuration unit uses a two-layer genetic algorithm to segment the operating time of the secondary water supply system into time periods, and configures corresponding water pump control schemes for each segmented operating period, wherein: The genetic algorithm in the first layer is used to segment the operating time of the secondary water supply system into time periods to obtain each operating time period; The genetic algorithm in the second layer is used to process and obtain the water pump control scheme with the lowest overall power consumption for the day in each of the running segments, so as to be the water pump control scheme configured in the running segment. The data required for the genetic algorithm in the second layer is the same data determined by the genetic algorithm in the first layer when performing time segmentation calculations.

2. The time-sharing pressure control system as described in claim 1, characterized in that, The water volume prediction unit includes: The first acquisition module is used to acquire export flow data and various influencing factor data from the original historical data; The first model training module, connected to the first acquisition module, is used to perform correlation analysis on the outlet flow data and the influencing factor data. Based on the analysis results, the first N outlet flow data and the multiple influencing factor data with correlation are used as the input data of the training sample, and the outlet flow data as the prediction result are used as the output data of the training sample to train the water volume prediction model. A water volume prediction model, connected to the first model training module, is used to input multiple outlet flow data and multiple related influencing factor data, which serve as the basis for prediction, into the trained water volume prediction model to obtain the water consumption prediction data.

3. The time-sharing pressure control system as described in claim 2, characterized in that, The water volume prediction unit also includes: The first lag analysis module is connected to the first acquisition module and the first model training module, respectively, and is used to analyze the lag data of the export flow data to the influencing factor data, and add the lag data to the input data of the training sample.

4. The time-sharing pressure control system as described in claim 1, characterized in that, The pressure prediction unit includes: The second acquisition module is used to acquire import pressure data, import and export flow data, and various influencing factor data from the original historical data; The second model training module, connected to the second acquisition module, is used to perform correlation analysis on the inlet pressure data, the inlet and outlet flow data, and the influencing factor data respectively. Based on the analysis results, the first N inlet pressure data, the multiple influencing factors with correlation, and the inlet and outlet flow data used as the prediction basis are used as the input data of the training sample, and the inlet pressure data used as the prediction result are used as the output data of the training sample to train the pressure prediction model. The pressure prediction model, connected to the second model training module, is used to input multiple import pressure data, as well as multiple related influencing factor data and import / export flow data, which serve as the basis for prediction, into the trained pressure prediction model to obtain the import pressure prediction data.

5. The time-sharing pressure control system as described in claim 4, characterized in that, The pressure prediction unit also includes: The second lag analysis module is connected to the second acquisition module and the second model training module, respectively. It is used to analyze the lag data of the import pressure data on the influencing factor data and the import and export flow data, and add the lag data to the input data of the training sample.

6. The time-sharing pressure control system as described in claim 1, characterized in that, The water pump fitting unit includes: The fourth acquisition module is used to acquire outlet flow rate data, inlet and outlet pressure data, power consumption data and pump operating frequency data from the original historical data, and to acquire pump factory parameter data from the pump-related original data. The first processing module, connected to the fourth acquisition module, is used to process the original historical data and the pump-related original data to obtain multiple standard relationship change fitting curves of the water pump under the power frequency. The second processing module is connected to the first processing module and the fourth acquisition module respectively. Based on the standard relationship change fitting curve, the original historical data and the pump-related original data, it processes the relationship change fitting curve between the flow rate, head, power and efficiency of the water pump at different operating frequencies, and uses it as the second correspondence.

7. The time-sharing pressure control system as described in claim 6, characterized in that, The standard relationship variation fitting curve includes: Fitted curve of the first standard relationship between flow rate and head; The fitted curve of the second standard relationship between flow rate and power; and The fitting curve for the change in the third standard relationship between flow rate and efficiency.

8. A time-segmented pressure control method applied to a secondary water supply system; characterized in that, Applied to the time-sharing pressure control system as described in any one of claims 1-7, and comprising: Step S1: Obtain the original historical data of the secondary water supply system and the original data related to each water pump. Step S2: Construct a water volume prediction model based on the original historical data, predict water volume prediction data based on the water volume prediction model, construct a pressure prediction model based on the original historical data, predict inlet pressure prediction data based on the pressure prediction model, fit a first correspondence between outlet flow rate and outlet pressure based on the original historical data, and fit a second correspondence between outlet flow rate and actual operating parameters of the pump based on the original historical data and the pump-related original data. Step S3: Based on the daily flow prediction data obtained from the original historical data, the original pump-related data, the water consumption prediction data, the inlet pressure prediction data, the first correspondence relationship, and the second correspondence relationship, the operating time of the secondary water supply system is divided into time periods, and a corresponding water pump control scheme is configured for each segmented operating period.

Images