Dynamic determination method for minimum night water consumption and leakage flow of independent metering area
By establishing the relationship between MNC and the number of households and data acquisition cycles, and dynamically determining MNC, the problem of large leakage flow error in the existing technology is solved, and more accurate leakage flow calculation and resource optimization are achieved.
Patent Information
- Application Number
- CN202510485871.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-17
- Publication Date
- 2025-08-19
AI Technical Summary
In the prior art, the method of determining the minimum night water consumption (MNC) in the independent metering area fails to fully consider the impact of the number of households and data collection period, resulting in large quantization errors of leakage flow.
By determining the relationship between MNC and the number of households and data acquisition period, an MNC value data set was constructed, and statistical methods were used to find the closest MNC value, and the MNC was dynamically determined to deduct the minimum night flow (MNF) to obtain the missing flow.
It improves the accuracy of MNC determination, improves the evaluation accuracy of DMA leakage flow, optimizes the allocation of pipeline network leakage control management resources, and reduces the cost of leakage control management.
Smart Images
Figure CN120509580A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of smart water management and water supply network leakage control, and specifically to a method for determining minimum nighttime water consumption (MNC) and its application in DMA leakage calculation. Background Art
[0002] In the field of water supply network leakage control, District Metered Area (DMA) technology, which divides the network into independent metering units with a size of no more than 5,000 households, enables continuous monitoring and leakage analysis of regional inlet flows. Minimum Night Flow (MNF), a core evaluation metric, is widely used to characterize DMA leakage levels by extracting the minimum inlet flow during low-water hours at night (typically 2:00-4:00 AM).
[0003] Based on the principle of conservation of mass, the true DMA loss flow should theoretically be calculated by deducting the minimum nighttime consumption (MNC) from the minimum nighttime flow (MNF). Due to the lack of an accurate method for determining the MNC, the MNF is often used directly to replace the DMA loss flow, assuming the nighttime MNC is zero. This approach can lead to an overestimation of the loss flow because residents are likely to use water at night (for example, by flushing the toilet), meaning the MNC should be greater than zero.
[0004] Another method uses a fixed empirical value, such as 2 L / h / household, to estimate the MNC. The MNC multiplied by the number of DMA households is subtracted from the MNF to obtain the lost flow rate. This empirical value can be derived from literature or from direct monitoring of nighttime water consumption by consumers. This monitoring method involves recording consumer water consumption at regular intervals (e.g., 30 minutes). After monitoring the consumption of multiple consumers over multiple days, the average nighttime water consumption is calculated and converted into flow rate, which is considered the MNC.
[0005] The disadvantage of the existing technology is that MNC does not consider the impact of the number of users and the data collection cycle, resulting in large errors.
[0006] The reason for this error is that residents' nighttime water use is a low-probability, random event. When the number of households is small (for example, considering only one household), there will always be periods when no residents use water, in which case the MNC should be equal to 0. However, when the number of households is large (for example, over 500 households), residents may use water at any time, in which case the MNC should be greater than 0. Regarding the data collection period, when the data collection period is short (for example, 5 minutes), the probability of water use is low, while when the data collection period is long (for example, 60 minutes), the probability of water use is high, which also leads to different MNCs. Therefore, both the number of households and the data collection period affect the MNC. Existing technologies do not consider this and simply use fixed empirical values, which leads to errors. Summary of the Invention
[0007] In light of this, the present invention provides a method for dynamically determining the minimum nighttime water consumption and loss flow in an independently metered area. This method, a dynamic MNC determination method, addresses the problem of inaccurate MNC estimation, which can lead to large errors in DMA loss flow quantification, by clarifying the relationship between MNC, household number, and data collection period.
[0008] The technical solution of the present invention is:
[0009] A method for dynamically determining the minimum nighttime water consumption of an independent metering area (DMA) comprises the following steps: 1) determining the relationship between the minimum nighttime water consumption (MNC) of an independent metering area (DMA) and the number of households and water consumption data collection period within the DMA; 2) using the relationship to construct a dataset of MNC values corresponding to different numbers of households and different data collection periods; and 3) searching for the closest MNC value in the dataset of MNC values based on the number of households and water consumption data collection period within the DMA to be analyzed.
[0010] Furthermore, the relationship between the minimum nighttime water consumption MNC of the independent metering area DMA and the number of households in the DMA and the water consumption data collection period is determined by statistical means.
[0011] Furthermore, step 1) specifically includes a data preparation stage and a data analysis stage, wherein the data preparation stage includes, within a data collection period of T, randomly extracting s household daily data, calculating the average water consumption corresponding to each time period, and taking the minimum value of the average water consumption corresponding to each time period as a single measurement value of the minimum nighttime water consumption corresponding to the data collection period of T and the number of households; randomly extracting s household daily data n times, and obtaining n measurement values of the minimum nighttime water consumption corresponding to the data collection period of T and the number of households; the data analysis stage includes statistically analyzing the n measurement values to determine the distribution of the minimum nighttime water consumption corresponding to the data collection period of T and the number of households; determining the relationship between the minimum nighttime water consumption and the number of households by keeping the data collection period unchanged and only changing the number of households; determining the relationship between the minimum nighttime water consumption and the data collection period by only changing the data collection period and keeping the number of households unchanged, wherein the daily data of one household refers to the water consumption data of one household at all sampling time points in one day.
[0012] Furthermore, in the data preparation stage, in order to increase the data capacity, a time aggregation method is used for the household daily data, and the data corresponding to the long data collection period under the multiple relationship is obtained from the data measured under the short data collection period.
[0013] Furthermore, when the MNC value data set does not contain the number of households x in the DMA to be analyzed, but only has the data collection period T, the MNC values corresponding to the two households x1 and x2 adjacent to the number of households x in the data collection period T are selected and recorded as MNC1 and MNC2; the weighted average of the MNC1 and MNC2 is performed to obtain the MNC of the number of households x in the DMA to be analyzed in the data collection period T, wherein the MNC in the weighted average is i The corresponding weighting coefficients, i=1, 2, and the number of households x in the DMA to be analyzed to the number of households x in the MNC value data set i is inversely proportional to the distance.
[0014] Furthermore, after obtaining the MNC values corresponding to different numbers of households and data collection periods, a statistical regression method is used to obtain the functional relationship between MNC, number of households, and data collection period; based on the functional relationship, the MNC value under any number of households and data collection period can be obtained.
[0015] A method for dynamically determining the leakage flow of an independent metering area is provided. Based on the method for dynamically determining the minimum nighttime water consumption of an independent metering area, the minimum nighttime water consumption (MNC) is deducted from the minimum nighttime flow (MNF) to obtain the DMA leakage flow.
[0016] Beneficial effects
[0017] 1. This invention breaks through the static limitations of the traditional fixed empirical value method, fully considers the randomness of residents' nighttime water consumption, determines the dynamic changes of MNC and DMA scale and data collection cycle, and improves the accuracy of MNC determination.
[0018] 2. The present invention improves the accuracy of DMA leakage flow determination and enhances the accuracy of pipeline network leakage assessment.
[0019] 3. Traditional technologies overestimate leakage flow, leading to excessive maintenance, or underestimate leakage, leading to delayed repairs, resulting in a mismatch of network maintenance resources. This invention optimizes resource allocation and reduces leakage control management costs by revealing the inherent relationship between MNC, household number, and data collection cycle.
[0020] 4. This invention promotes the upgrade of loss detection technology from experience-driven to model-driven by quantifying the impact of user scale and time resolution on MNC. BRIEF DESCRIPTION OF THE DRAWINGS
[0021] Figure 1 , a specific flow chart of an embodiment of the present invention:
[0022] Figure 2 , the change of MNC with size(s);
[0023] Figure 3 , flow chart of the method of the present invention. DETAILED DESCRIPTION
[0024] The technical solutions of the present invention are as follows:
[0025] (1) Collect water consumption data from smart water meters of multiple households with a data collection period of T for multiple days. The number of households should be no less than 2,000, the number of days should be no less than 30 days, T should be no more than 30 minutes, and the data accuracy should be 1L;
[0026] (2) Align the T-minute water consumption data of different users on different dates collected in step (1) by timestamp to generate a structured dataset;
[0027] (3) Define one user's data for one day as one household-day data, and record the data volume of the data set as S household-days. According to the parameters recommended in step (1), S is not less than 60,000;
[0028] (4) Randomly select s household daily data from the data set, calculate the average water consumption data corresponding to each period, and take the minimum value, recorded as MNC(s);
[0029] (5) Repeat step (4) n times to obtain n MNC(s), where n is preferably not less than 100;
[0030] (6) Calculate the mean and standard deviation of n MNC(s) and fit the distribution of MNC(s) using normal distribution;
[0031] (7) Preset s = {1, 10, 50, 100, 200, 500, 1000, 2000, 3000, 4000, 5000, 6000} household days, take different values of s, repeat steps (4)-(6) to obtain the mean, standard deviation and distribution of MNC(s) corresponding to different values of s;
[0032] (8) Aggregate the data collected in step (1) into feasible data collection periods to generate a new data set. The data collection period should be selected from the currently common 5 min, 10 min, 15 min, 30 min, 60 min, etc.
[0033] (9) For each data collection period, execute steps (4)-(8) to obtain the mean, standard deviation, and distribution of MNC(s) corresponding to different data collection periods;
[0034] (10) Consider the two variables s and T and obtain the corresponding MNC(s,T);
[0035] (11) Collect the MNF of a certain DMA and find the closest MNC (s = N, T = TMNF) value based on the number of households N in the DMA and the data collection period TMNF of the MNF; the closest value refers to the MNC value when s is closest to N and T is closest to TMNF.
[0036] (12) Subtract the product of N and MNC (s = N, T = TMNF) from MNF to obtain the DMA leakage flow.
[0037] The above is only one implementation of the method and is not intended to limit the technical solution of the present invention.
[0038] In order to express this technical solution more clearly, the following is an explanation with reference to specific data.
[0039] This embodiment collects water consumption data of 2,000 users every 30 minutes for 300 days. The data accuracy is 1L, so the total data volume is 600,000 household days, and each household day data includes 48 water consumption data.
[0040] After aligning all the data by timestamp, we construct the following dataset: t1-t48 represents different times on the same household day, and s1-s600000 represents different household days. This dataset consists of 600,000 rows and 48 columns of water consumption data.
[0041] Table 1 Original dataset of water consumption data
[0042] t1 t2 t3 …… t47 t48 s1 0 0 7 …… 12 0 s2 0 5 0 …… 0 8 …… …… …… …… …… …… …… s599990 6 0 …… 10 0 s600000 0 0 3 …… 0 0
[0043] Assume s = {1, 10, 50, 100, 200, 500, 1000, 2000, 3000, 4000, 5000, 6000} household-days. For each value of s, randomly sample 100 times to obtain the sampled data. For example, with s = 10, 10 household-days are randomly sampled from the original dataset each time. Assume that the result of one of these samplings is as shown in Table 2. The average of each row in Table 2 is calculated by column, and the result is shown in the last row of Table 2.
[0044] Table 2 Random sampling results when s=10
[0045]
[0046]
[0047] It can be seen that among the time periods in Table 2, the time period with the lowest average water consumption is t2, with a water consumption of 0.07 L / household. Since the data collection cycle is 30 minutes, after converting the unit of water consumption to L / h / household, the value is 0.14 L / h / household. That is, when s = 10, the MNC (s = 10) of this sampling is 0.14 L / h / household.
[0048] The random sampling was repeated 100 times to obtain 100 MNC (s=10) values, and the mean and standard deviation were calculated.
[0049] For each value of s in the household-day range s = {1, 10, 50, 100, 200, 500, 1000, 2000, 3000, 4000, 5000, 6000}, the above random sampling and statistical analysis of the mean and standard deviation of MNC are repeated to obtain the mean and standard deviation of MNC for each value of s as shown in Table 3.
[0050] Table 3 Mean and standard deviation of MNC corresponding to different s values
[0051]
[0052]
[0053] Statistically calculate the MNC distribution obtained by 100 random samplings for each s value, and plot the MNC distribution as s changes, as shown in the figure below: Figure 2As shown in Figure 3, the mean MNC gradually increases with increasing s, and the consistency of its distribution becomes increasingly better. The KS test shows that the p-values are greater than 0.5 for s = 4000 and s = 5000, indicating that the two have identical distributions. Table 3 also shows that when s reaches 4000 household-days, the mean MNC and its distribution no longer change significantly. If MNC still changes significantly with s, then s should be increased further and the aforementioned random sampling analysis should be repeated.
[0054] The above steps complete the MNC analysis process for a 30-minute sampling period. Next, the original data is aggregated into data with a 60-minute sampling period. The aggregation method is to add the data corresponding to t1 and t2, the data corresponding to t3 and t4, and so on, adding the data corresponding to t47 and t48. The original 48-period data for a household per day is aggregated into 24 periods, forming a new data set, as shown in Table 4.
[0055] Table 4 Dataset after the data collection period is aggregated from 30 minutes to 60 minutes
[0056] t1 t2 …… t24 s1 0 0 …… 12 s2 5 0 …… 8 …… …… …… …… …… s599990 6 0 …… 10 s600000 0 3 …… 0
[0057] Following the aforementioned analysis steps for the change of MNC with s, the change of MNC was recalculated for the data set after it was aggregated into 60-minute data.
[0058] The results before and after aggregation are combined to construct a table showing how MNC changes with scale s and data collection period T, as shown in Table 5.
[0059] Table 5. Mean MNC values for different scales s and different data collection periods T.
[0060] s T=30min T=60min 1 0.01 0.02 10 0.06 0.17 50 0.26 0.49 100 0.40 0.58 200 0.51 0.64 500 0.60 0.70 1000 0.65 0.73 2000 0.68 0.74 3000 0.70 0.74 4000 0.71 0.75 5000 0.71 0.75 6000 0.71 0.75
[0061] As can be seen from Table 5, MNC varies significantly with DMA scale and data collection period. Therefore, using fixed empirical values in existing techniques can introduce significant errors. It should be noted that because the original data collection period in this example is 30 minutes, aggregation can only be performed with a longer collection period, such as 60 minutes. If the original data collection period is 5 minutes, then aggregation can be performed with periods of 10, 15, 30, or 60 minutes.
[0062] Collect the MNF value of a DMA, assuming its value is 3m 3 / h, the data collection cycle is 30min, the number of DMA households is 1600, and the closest data corresponding to Table 2 is 0.68L / h / household when T=30min and s=2000. The leakage flow of this DMA is 3m 3 / h-0.68L / h / household×1600 households=1.912m 3 / h
[0063] Furthermore, when looking up the table to obtain the MNC value, this solution uses the closest value. Alternatively, a weighted average of two adjacent values can be used. For example, if the DMA has 1600 households, when taking the weighted average of the MNCs corresponding to s = 1000 and s = 2000, the weight is determined by the difference between the DMA number of households and the s value; the larger the difference, the smaller the weight. In this example, the difference between the DMA number of households and s = 1000 and s = 2000 is 600 and 400, respectively. Therefore, the weights are 400 / (600 + 400) = 0.4 and 600 / (600 + 400) = 0.6, respectively. The corresponding MNC value for this DMA is MNC(s = 1000) × 0.4 + MNC(s = 2000) × 0.6 = 0.65 × 0.4 + 0.68 × 0.6 = 0.668 L / h / household.
[0064] In addition, after obtaining the MNC values corresponding to different s and T, the statistical regression method can also be used to obtain the functional relationship between MNC, s and T. For any DMA, the number of households N and the MNF data collection period T are MNF Substituting s and T respectively, we can get the MNC value.
Claims
1. A method for dynamically determining the minimum nighttime water consumption in an independent metering area, characterized by: Step 1) Determine the relationship between the minimum nighttime water consumption MNC of the independent metering area DMA and the number of households in the DMA and the water consumption data collection period; Step 2) using the relationship, constructing MNC value datasets corresponding to different numbers of households and different data collection periods; Step 3) According to the number of households in the DMA to be analyzed and the water consumption data collection period, the closest MNC value is found in the MNC value data set.
2. The method for dynamically determining the minimum nighttime water consumption in an independent metering area according to claim 1, characterized in that: The relationship between the minimum nighttime water consumption MNC of the independent metering area DMA and the number of households in the DMA and the water consumption data collection period is determined by statistical methods.
3. The method for dynamically determining the minimum nighttime water consumption in an independent metering area according to claim 2, characterized in that: Step 1) specifically includes a data preparation stage and a data analysis stage, wherein the data preparation stage includes, in a data collection period of T, randomly extracting s household daily data, calculating the average water consumption corresponding to each time period, and taking the minimum value of the average water consumption corresponding to each time period as a single measurement value of the minimum nighttime water consumption corresponding to the data collection period of T and the number of households; randomly extracting s household daily data n times, and obtaining n measurement values of the minimum nighttime water consumption corresponding to the data collection period of T and the number of households; the data analysis stage includes statistically analyzing the n measurement values to determine the distribution of the minimum nighttime water consumption corresponding to the data collection period of T and the number of households; determining the relationship between the minimum nighttime water consumption and the number of households by keeping the data collection period unchanged and only changing the number of households; determining the relationship between the minimum nighttime water consumption and the data collection period by only changing the data collection period and keeping the number of households unchanged, wherein the daily data of one household refers to the water consumption data of one household at all sampling time points in one day.
4. The method for dynamically determining the minimum nighttime water consumption in an independent metering area according to claim 3, characterized in that: In the data preparation stage, in order to increase the data capacity, a time aggregation method is used for the household daily data, and the data corresponding to the long data collection period under the multiple relationship is obtained from the data measured under the short data collection period.
5. A method for dynamically determining the minimum nighttime water consumption in an independent metering area according to any one of claims 1 to 4, characterized in that: When the MNC value data set does not contain the number of households x in the DMA to be analyzed, but only has the data collection period T, the MNC values corresponding to the two households x1 and x2 adjacent to the number of households x in the data collection period T are selected and recorded as MNC1 and MNC2; the weighted average of the MNC1 and MNC2 is performed to obtain the MNC of the number of households x in the DMA to be analyzed in the data collection period T, wherein the MNC in the weighted average is i The corresponding weighting coefficients, i=1, 2, and the number of households x in the DMA to be analyzed to the number of households x in the MNC value data set i is inversely proportional to the distance.
6. A method for dynamically determining the minimum nighttime water consumption in an independent metering area according to any one of claims 1 to 4, characterized in that: After obtaining the MNC values corresponding to different numbers of households and data collection periods, a statistical regression method is used to obtain the functional relationship between MNC, number of households, and data collection period. Based on the functional relationship, the MNC value for any number of households and data collection period can be obtained.
7. A method for dynamically determining the leakage flow in an independent metering area, based on the method according to any one of claims 1 to 6, characterized in that: The DMA leakage flow is obtained by deducting the minimum nighttime water consumption MNC from the minimum nighttime flow MNF.