Metering asset intelligent demand prediction system and method based on algorithm

Abstract

The invention discloses a metering asset intelligent demand prediction system and method based on an algorithm, and relates to the technical field of metering assets, and the method can gradually output a comprehensive adjustment demand prediction value and a final demand prediction value through introducing a basic demand prediction model and a progressive algorithm, and can accurately reflect the future equipment demand condition. The system establishes a multi-level early warning mode according to the current stock S (t) and the early warning threshold L, so that the early warning level can be dynamically adjusted according to different states of the stock. And in combination with the calculation of the supply and demand difference E1, the system can be helped to carry out preliminary evaluation, so that inventory and demands can be meticulously analyzed in a prediction stage. Compared with a traditional static threshold value monitoring method, the intelligent early warning mechanism has higher response speed and accuracy, the situation of stock overstock or supply interruption is avoided, and therefore reasonable storage of metering assets is guaranteed.

Description

Technical Field

[0001] The present invention belongs to the technical field of metering assets, and in particular relates to an algorithm-based intelligent demand forecasting system and method for metering assets. Background Art

[0002] In the power industry, metering assets include energy meters, transformers and other equipment. The management and demand forecasting of these equipment are crucial to the operational efficiency and cost control of the power grid. Through this method, data from power grid-related systems can be collected and analyzed in real time, so as to make intelligent forecasts of equipment demand, and monitor inventory status through a multi-level early warning system to ensure a smooth and efficient supply of equipment.

[0003] At present, the main problem facing metering asset management is the lack of accuracy in inventory management and demand forecasting. Due to the complex changes in equipment demand and the influence of many factors, such as grid load fluctuations, equipment aging and environmental conditions, traditional management methods rely on experience or static analysis, which is difficult to cope with the dynamic changes in actual demand. This often leads to too much or too little equipment inventory. Inventory backlogs will increase storage costs and maintenance pressure, while insufficient inventory will cause supply chain disruptions and affect the normal supply of power equipment. In addition, the early warning mechanism of the existing system is not perfect and cannot respond to fluctuations in equipment demand in a timely manner, resulting in low operational efficiency. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention provides an algorithm-based intelligent demand forecasting system and method for metering assets, which solves the problems mentioned in the background technology.

[0005] To achieve the above objectives, the present invention is implemented by the following technical scheme: comprising the following steps:

[0006] S1. Build a metering asset management and control platform and set up an API application program interface to connect with the relevant systems of metering assets, collect metering asset demand data in relevant systems in real time, pre-process the metering asset demand data, integrate it into a demand data set, and then set up a time series database to store the demand data set;

[0007] S2. Construct a basic demand forecasting algorithm model, perform preliminary calculations to output a basic demand forecast value Db(t), and use a progressive algorithm based on the basic demand forecast value Db(t) to sequentially output a comprehensive adjustment demand forecast value Dadj(t) and a final demand forecast value Dh(t);

[0008] S3. A multi-level early warning mode is established based on the current inventory S(t) of the measurement asset and the inventory threshold L to obtain the early warning level Pw(t). The difference between the current inventory S(t) of the measurement asset and the obtained final demand forecast value Dh(t) is calculated to obtain the supply-demand difference value E1. Based on the output of the supply-demand difference value, a preliminary evaluation and analysis of the measurement assets and inventory in the future forecast stage is performed.

[0009] S4. When the preliminary evaluation result is abnormal, the optimized demand forecast value Dopt(t) is calculated and output based on the current use of the power grid equipment, and the difference between the current inventory of the metering asset S(t) and the optimized demand forecast value Dopt(t) is calculated to obtain the second supply and demand difference value E2;

[0010] S5. Based on the second supply-demand difference value E2, a second evaluation is performed on the demand for measurement assets after optimization, and relevant information prompts are generated through the measurement asset management and control platform.

[0011] Preferably, S1 includes S11, S12 and S13;

[0012] S11. Build a metering asset management and control platform, set up an API application program interface to connect with the relevant systems of metering assets, and set up to collect metering asset demand data from relevant systems every hour;

[0013] Related systems include storage systems, power grid systems, and meteorological systems;

[0014] The measurement asset demand data includes inventory S, demand change rate Dr, measurement asset equipment storage age Ts, warehouse temperature Hw, warehouse humidity Hs, failure rate Fr, power grid fluctuation rate Ve and environmental disaster index Wd;

[0015] The inventory quantity S and the inventory age Ts of the measuring asset equipment are collected and obtained through the warehousing system;

[0016] The demand change rate Dr is calculated using the historical outbound data provided by the warehousing system and the time series analysis algorithm to output the demand change rate Dr of the metering asset;

[0017] The warehouse temperature Hw is obtained by collecting the temperature parameters in the warehouse through the temperature sensor integrated in the storage system in real time;

[0018] The humidity Hs of the warehouse is obtained by collecting humidity parameters in the warehouse through the humidity sensor integrated in the warehouse system in real time;

[0019] The failure rate Fr is calculated by collecting the historical use and maintenance records of the power grid system. The specific algorithm formula is: In the formula, Nf represents the number of equipment failures per unit time, and Tuse represents the cumulative operating time of the equipment;

[0020] The grid fluctuation rate Ve is calculated by collecting the power load P(t) at time t and the reference load value Pavg from the grid system. The specific algorithm formula is: Where T represents the total time step;

[0021] The environmental disaster index Wd is collected through the meteorological system.

[0022] Preferably, S12, preprocessing the collected metrology asset demand data in the metrology asset management and control platform, wherein the preprocessing method includes marking timestamps and standardization processing, and after the preprocessing, integrating and integrating to obtain the demand data set;

[0023] The marking timestamp is used to mark each parameter in the collected metering asset demand data according to the collection time, and obtain the inventory S(t), demand change rate Dr(t), metering asset equipment age Ts(t), warehouse temperature Hw(t), warehouse humidity Hs(t), failure rate Fr(t), power grid fluctuation rate Ve(t) and environmental disaster index Wd(t) at time t respectively;

[0024] S13. By constructing a time series database and setting a write port, a write port, a real-time data table and a historical data table, the demand data set obtained after preprocessing is stored in the real-time data table through the write port, and the demand data set stored last time is automatically transferred to the historical data table and stored in chronological order.

[0025] Preferably, S2 includes S21, S22 and S23;

[0026] S21, extract the demand data set in the real-time data table through the write port, input it into the basic demand forecasting algorithm model, calculate and output the basic demand forecast value Db(t), and forecast the demand of the measurement asset at the next time t;

[0027] The basic demand forecast value Db(t) is calculated and output by the following basic demand forecast algorithm model;

[0028] Db(t)=a1·H(t)+a2·S(t)+a3·Ve(t)+a4·Wd(t);

[0029] Where H(t) represents the usage frequency of metering assets at historical time t, a1, a2, a3 and a4 represent the preset weight values ​​of the historical metering asset usage frequency H(t), inventory S(t), grid volatility Ve(t) and environmental disaster index Wd(t), respectively, and a1+a2+a3+a4=1, and the specific values ​​are set by the user.

[0030] Preferably, S22, based on the obtained basic demand forecast value Db(t), a progressive algorithm is used to calculate and output a comprehensive adjusted demand forecast value Dadj(t) in combination with the demand change rate Dr(t) and the metering asset equipment storage age Ts(t), and then a final demand forecast value Dh(t) is obtained by performing a progressive calculation based on the comprehensive adjusted demand forecast value Dadj(t) in combination with the warehouse temperature Hw(t) and the warehouse humidity Hs(t);

[0031] The comprehensive adjustment demand forecast value Dadj(t) is further calculated and obtained through the basic demand forecast value Db(t);

[0032] The final demand forecast value Dh(t) is further calculated and obtained by comprehensively adjusting the demand forecast value Dadj(t);

[0033] The comprehensive adjustment demand forecast value Dadj(t) and the final demand forecast value Dh(t) are calculated and obtained by the following progressive algorithm;

[0034]

[0035] In the formula, represents the differential variable of the demand change rate, d represents the differential function, dt represents the differential variable of time t, η represents the temperature influence adjustment coefficient, μ represents the humidity influence adjustment coefficient, Hwopt represents the ideal temperature condition, Hsopt represents the ideal humidity condition, δ represents the adjustment coefficient of the demand change rate, ε represents the adjustment coefficient of the inventory age of the metering asset equipment, and ζ represents the adjustment coefficient of the failure rate.

[0036] Preferably, S3 includes S31, S32 and S33;

[0037] S31. Establish a multi-level warning mode according to the current inventory quantity S(t) and the inventory threshold L to obtain the warning level Pw(t). The inventory threshold L includes the inventory upper limit threshold Lmax and the inventory lower limit threshold Lmin.

[0038] The warning level Pw(t) is obtained by the multi-level warning mode;

[0039]

[0040] In the formula, 0 represents the first-level inventory warning, 1 represents the second-level inventory warning, and 2 represents the third-level inventory warning;

[0041] Level 1 inventory warning > Level 2 inventory warning > Level 3 inventory warning;

[0042] Level 1 inventory warning means that the inventory is in the normal range. At this time, it is considered that the inventory is sufficient to meet future demand and no immediate replenishment or adjustment measures are required;

[0043] Level 2 inventory warning means that the inventory level is close to the minimum safety stock threshold. Although the inventory level is still within an acceptable range, it is close to the warning state. It is necessary to pay attention to the inventory consumption and may need to take replenishment actions in the near future.

[0044] The third-level inventory warning means that the inventory level is lower than the minimum safety threshold. At this time, the inventory has entered an emergency state. The inventory level is insufficient to meet demand, which may cause a supply chain disruption. Immediate replenishment measures must be taken to avoid out-of-stock problems.

[0045] Preferably, S32, calculating the inventory difference by using the predicted final demand forecast value Dh(t) and the current inventory quantity S(t) to obtain the supply-demand difference value E1;

[0046] The supply and demand difference value E1 is calculated and output by the following difference calculation formula;

[0047] E1=S(t)-Dh(t);

[0048] S33. Based on the output result of the obtained supply-demand difference value E1, a preliminary comparative evaluation is performed to analyze the inventory status of the current measurement assets. The specific evaluation contents are as follows;

[0049] When the supply-demand difference value E1>0, it means that the inventory of the predicted metering asset equipment is sufficient, and there is no need to send an early warning;

[0050] When the supply-demand difference value E1≤0, it means that the predicted inventory of metering asset equipment is insufficient, and secondary optimization is performed at this time.

[0051] Preferably, S4 includes S41 and S42;

[0052] S41. When the preliminary comparative evaluation result shows that the inventory is insufficient or abnormal, the optimized demand forecast value Dopt(t) is further calculated based on the equipment status of the metering assets and the final demand forecast value Dh(t) to optimize the metering assets;

[0053] The optimized demand forecast value Dopt(t) is calculated and output by the following algorithm formula;

[0054]

[0055] In the formula, κ represents the adjustment coefficient of failure rate to forecast demand, λ represents the adjustment coefficient of inventory to forecast demand, Fmax represents the upper limit of failure rate, and Smax represents the upper limit of inventory;

[0056] S42, performing a secondary difference calculation based on the obtained optimized demand forecast value Dopt(t) and the current inventory S(t) of the metered asset to obtain a second supply-demand difference value E2;

[0057] The second supply-demand difference value E2 is calculated by the following difference calculation formula:

[0058] E2 = S(t) - Dopt(t).

[0059] Preferably, S5 includes S51;

[0060] S51. Based on the obtained second supply-demand difference value E2 output result, and in combination with the multi-level early warning mode, a secondary comparative evaluation is performed to analyze the inventory status of the final measurement asset equipment. The specific evaluation contents are as follows;

[0061] When the second supply-demand difference value E2>0, it means that the inventory of the measurement asset equipment is predicted to be sufficient. At this time, a first-level inventory warning is sent through the measurement asset management and control platform;

[0062] When the second supply-demand difference value E2≤0, it means that the inventory of the predicted measurement asset equipment is insufficient. At this time, a secondary inventory warning is sent through the measurement asset management and control platform;

[0063] When the second supply-demand difference value E2≤2, it indicates that the predicted inventory of metering asset equipment is abnormal. At this time, a third-level inventory warning is sent through the metering asset management and control platform.

[0064] An algorithm-based intelligent demand forecasting system for metering assets, comprising a demand collection module, a demand forecasting algorithm module, a multi-level early warning and preliminary analysis module, a secondary forecasting optimization analysis module and a multi-level early warning combination module;

[0065] The demand collection module builds a metering asset management and control platform and sets up an API application program interface to connect with the relevant systems of metering assets, collects metering asset demand data in relevant systems in real time, pre-processes the metering asset demand data, integrates it into a demand data set, and then sets up a time series database to store the demand data set;

[0066] The demand forecasting algorithm module constructs a basic demand forecasting algorithm model, performs preliminary calculations and outputs the basic demand forecast value Db(t), and uses a progressive algorithm based on the basic demand forecast value Db(t) to sequentially output the comprehensive adjustment demand forecast value Dadj(t) and the final demand forecast value Dh(t);

[0067] The multi-level warning and preliminary analysis module establishes a multi-level warning mode based on the current inventory of the metering asset S(t) and the inventory threshold L, obtains the warning level Pw(t), and then calculates the difference between the current inventory of the metering asset S(t) and the obtained final demand forecast value Dh(t), obtains the supply-demand difference value E1, and based on the output of the supply-demand difference value, conducts a preliminary evaluation and analysis of the metering assets and inventory in the future forecast stage;

[0068] The secondary forecast optimization analysis module calculates and outputs the optimized demand forecast value Dopt(t) based on the abnormal result of the preliminary evaluation, and calculates the difference between the current inventory of the metering asset S(t) and the optimized demand forecast value Dopt(t) to obtain the second supply and demand difference value E2;

[0069] The multi-level early warning combination module performs a secondary evaluation based on the second supply-demand difference value E2 to optimize the demand for metering assets, and generates relevant information prompts through the metering asset management and control platform.

[0070] Compared with the prior art, the beneficial effects are:

[0071] (1) The metering asset management and control platform constructed by this method realizes the real-time collection and preprocessing of metering asset demand data, and stores these data sets through a time series database to ensure the integrity and timeliness of the data. Based on this platform, multi-dimensional data from the warehousing system, power grid system and meteorological system can be integrated to make the data richer and more accurate. The API interface in the system can ensure efficient connection and data transmission between different systems, and automatically complete data collection and prediction tasks. Compared with traditional manual processing or static analysis, this solution can capture the dynamic changes of data in a timely manner, ensure the accuracy of metering asset demand data, and effectively reduce the impact of information lag.

[0072] (2) By introducing the basic demand forecasting model and the progressive algorithm, this method can gradually output the comprehensive adjustment demand forecast value and the final demand forecast value, accurately reflecting the future equipment demand situation. The system establishes a multi-level warning mode based on the current inventory level S(t) and the warning threshold L, so that the warning level can be dynamically adjusted according to the different states of the inventory level. Combined with the calculation of the supply and demand difference E1, it can help the system to make a preliminary assessment, so that the inventory and demand can be analyzed in detail in the forecasting stage. Compared with the traditional static threshold monitoring method, this intelligent warning mechanism has a higher response speed and accuracy, avoids inventory backlogs or supply interruptions, and thus ensures the reasonable reserve of metering assets.

[0073] (3) When the supply-demand difference analysis indicates that the inventory is insufficient or abnormal, the system further calculates the optimized demand forecast value Dopt(t) by combining the usage of the power grid equipment, and obtains the second supply-demand difference value E2 through the secondary optimization algorithm. This secondary optimization mechanism can further consider the impact of uncommon parameters such as equipment failure rate and power grid fluctuation rate on demand on the basis of the initial forecast, so as to achieve more accurate forecasting and demand adjustment. At this time, the system can send different levels of early warning prompts in time according to the changes in inventory, thereby avoiding supply chain problems caused by insufficient inventory. The optimized demand forecast not only improves the accuracy of supply and demand balance, but also enhances the dynamic response capability of the system, making equipment management more flexible and ensuring the continuity and safety of power grid operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] Figure 1 A schematic diagram of the steps of an algorithm-based intelligent demand forecasting method for metering assets of the present invention;

[0075] Figure 2 The present invention is a flowchart of an algorithm-based intelligent demand forecasting system for metering assets. DETAILED DESCRIPTION

[0076] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0077] Example 1

[0078] See also Figure 1 The present invention provides an algorithm-based intelligent demand forecasting method for metering assets. To achieve the above purpose, the present invention is implemented through the following technical solutions: comprising the following steps:

[0079] S1. Build a metering asset management and control platform and set up an API application program interface to connect with the relevant systems of metering assets, collect metering asset demand data in relevant systems in real time, pre-process the metering asset demand data, integrate it into a demand data set, and then set up a time series database to store the demand data set;

[0080] S2. Construct a basic demand forecasting algorithm model, perform preliminary calculations to output a basic demand forecast value Db(t), and use a progressive algorithm based on the basic demand forecast value Db(t) to sequentially output a comprehensive adjustment demand forecast value Dadj(t) and a final demand forecast value Dh(t);

[0081] S3. A multi-level early warning mode is established based on the current inventory S(t) of the measurement asset and the inventory threshold L to obtain the early warning level Pw(t). The difference between the current inventory S(t) of the measurement asset and the obtained final demand forecast value Dh(t) is calculated to obtain the supply-demand difference value E1. Based on the output of the supply-demand difference value, a preliminary evaluation and analysis of the measurement assets and inventory in the future forecast stage is performed.

[0082] S4. When the preliminary evaluation result is abnormal, the optimized demand forecast value Dopt(t) is calculated and output based on the current use of the power grid equipment, and the difference between the current inventory of the metering asset S(t) and the optimized demand forecast value Dopt(t) is calculated to obtain the second supply and demand difference value E2;

[0083] S5. Based on the second supply-demand difference value E2, a second evaluation is performed on the demand for measurement assets after optimization, and relevant information prompts are generated through the measurement asset management and control platform.

[0084] In this embodiment, the method can collect the demand data of metering assets in real time by constructing a metering asset management and control platform and setting an API interface to connect with relevant data. After preprocessing these data, they are stored in a time series database to ensure that the data is complete, reliable and timely. This process effectively solves the problems of data fragmentation and untimely collection in the traditional method. By constructing a basic demand forecasting algorithm model, a preliminary calculation is performed to output the basic demand forecast value Db(t), and a progressive algorithm is used to gradually calculate the comprehensive adjustment demand forecast value Dadj(t) and the final demand forecast value Dh(t), so as to realize a multi-stage and precise forecast of the demand for metering assets, overcoming the problems of inaccurate forecasting and reliance on experience in traditional methods. Combined with the current inventory S(t) and the inventory threshold L, a multi-level early warning mechanism is established, which can make a preliminary assessment of the inventory status through the supply and demand difference value E1, and obtain information on whether the inventory is sufficient in real time, thereby avoiding inventory backlog or supply interruption. When the preliminary assessment result is abnormal, the demand forecast value Dopt(t) is further optimized by combining the use of power grid equipment, and the second supply and demand difference value E2 is calculated to ensure that the forecast adjustment is more accurate and flexible to meet actual needs. Based on the optimized supply-demand difference value E2, a secondary evaluation is performed to generate relevant information prompts to help managers understand the optimized inventory status and demand adjustment plan in a timely manner, thereby improving intelligent response capabilities. Compared with the current traditional metering asset management technology, this method significantly improves the prediction accuracy and response speed by introducing real-time data collection, multi-dimensional data processing, progressive algorithms and multi-level early warning mechanisms. It avoids the problem of supply and demand imbalance in inventory management through timely identification and prediction optimization of abnormal situations, thereby improving resource utilization efficiency, reducing storage costs, and ensuring the supply security of power grid equipment.

[0085] Example 2

[0086] This embodiment is explained in Example 1, please refer to Figure 1 , specifically: S1 includes S11, S12 and S13;

[0087] S11. Build a metering asset management and control platform, set up an API application program interface to connect with the relevant systems of metering assets, and set up to collect metering asset demand data from relevant systems every hour;

[0088] Related systems include storage systems, power grid systems, and meteorological systems;

[0089] The measurement asset demand data includes inventory S, demand change rate Dr, measurement asset equipment storage age Ts, warehouse temperature Hw, warehouse humidity Hs, failure rate Fr, power grid fluctuation rate Ve and environmental disaster index Wd;

[0090] The inventory quantity S and the inventory age Ts of the measuring asset equipment are collected and obtained through the warehousing system;

[0091] The demand change rate Dr is calculated using the historical outbound data provided by the warehousing system and the time series analysis algorithm to output the demand change rate Dr of the metering asset;

[0092] The warehouse temperature Hw is obtained by collecting the temperature parameters in the warehouse through the temperature sensor integrated in the storage system in real time;

[0093] The humidity Hs of the warehouse is obtained by collecting humidity parameters in the warehouse through the humidity sensor integrated in the warehouse system in real time;

[0094] The failure rate Fr is calculated by collecting the historical use and maintenance records of the power grid system. The specific algorithm formula is: In the formula, Nf represents the number of equipment failures per unit time, and Tuse represents the cumulative operating time of the equipment;

[0095] The grid fluctuation rate Ve is calculated by collecting the power load P(t) at time t and the reference load value Pavg from the grid system. The specific algorithm formula is: Where, T represents the total time step;

[0096] The environmental disaster index Wd is collected through the meteorological system.

[0097] S12. Preprocessing the collected metrology asset demand data in the metrology asset management and control platform, the preprocessing method includes marking timestamps and standardization processing, and after preprocessing, integrating and integrating to obtain the demand data set;

[0098] The marking timestamp is used to mark each parameter in the collected metering asset demand data according to the collection time, and obtain the inventory S(t), demand change rate Dr(t), metering asset equipment age Ts(t), warehouse temperature Hw(t), warehouse humidity Hs(t), failure rate Fr(t), power grid fluctuation rate Ve(t) and environmental disaster index Wd(t) at time t respectively;

[0099] S13. By constructing a time series database and setting a write port, a write port, a real-time data table and a historical data table, the demand data set obtained after preprocessing is stored in the real-time data table through the write port, and the demand data set stored last time is automatically transferred to the historical data table and stored in chronological order.

[0100] In this embodiment, the method constructs a metering asset management and control platform and sets an API interface, so that the system can connect to related systems such as warehousing, power grid and meteorology in real time to collect key demand data on a regular basis. This real-time data collection method greatly improves the accuracy and timeliness of the data and solves the problem of delayed or incomplete data. By preprocessing the collected data, including marking timestamps and standardization, the temporal consistency and standardization of each parameter data are ensured. This method solves the inconsistency problem in traditional data processing and enhances the availability and accuracy of the data. Further, by constructing a time series database, the classified storage of data and the automatic transfer of historical data are realized, ensuring the systematicness and traceability of the data.

[0101] Example 3

[0102] This embodiment is explained in Example 2. Please refer to Figure 1 , specifically: S2 includes S21, S22 and S23;

[0103] S21, extract the demand data set in the real-time data table through the write port, input it into the basic demand forecasting algorithm model, calculate and output the basic demand forecast value Db(t), and forecast the demand of the measurement asset at the next time t;

[0104] The basic demand forecast value Db(t) is calculated and output by the following basic demand forecast algorithm model;

[0105] Db(t)=a1·H(t)+a2·S(t)+a3·Ve(t)+a4·Wd(t);

[0106] Where H(t) represents the usage frequency of metering assets at historical time t, a1, a2, a3 and a4 represent the preset weight values ​​of the historical metering asset usage frequency H(t), inventory S(t), grid volatility Ve(t) and environmental disaster index Wd(t), respectively, and a1+a2+a3+a4=1, and the specific values ​​are set by the user.

[0107] S22. Based on the obtained basic demand forecast value Db(t), a progressive algorithm is used to calculate and output the comprehensive adjusted demand forecast value Dadj(t) in combination with the demand change rate Dr(t) and the inventory age of the metering asset equipment Ts(t). Then, a progressive calculation is performed based on the comprehensive adjusted demand forecast value Dadj(t) in combination with the warehouse temperature Hw(t) and the warehouse humidity Hs(t) to obtain the final demand forecast value Dh(t).

[0108] The comprehensive adjustment demand forecast value Dadj(t) is further calculated and obtained through the basic demand forecast value Db(t);

[0109] The final demand forecast value Dh(t) is further calculated and obtained by comprehensively adjusting the demand forecast value Dadj(t);

[0110] The comprehensive adjustment demand forecast value Dadj(t) and the final demand forecast value Dh(t) are calculated and obtained by the following progressive algorithm;

[0111]

[0112] In the formula, represents the differential variable of the demand change rate, d represents the differential function, dt represents the differential variable of time t, η represents the temperature influence adjustment coefficient, μ represents the humidity influence adjustment coefficient, Hwopt represents the ideal temperature condition, Hsopt represents the ideal humidity condition, δ represents the adjustment coefficient of the demand change rate, ε represents the adjustment coefficient of the inventory age of the metering asset equipment, and ζ represents the adjustment coefficient of the failure rate.

[0113] In this embodiment, the method extracts the demand data set in the real-time data table, inputs these data into the basic demand forecasting algorithm model, calculates the basic demand forecast value Db(t), and provides a preliminary estimate for the metering asset demand at a certain moment in the future. This process forms a more comprehensive demand forecasting model by comprehensively considering key parameters such as historical usage frequency, inventory, power grid volatility, and environmental disaster index, solving the problems of single data dimension and insufficient prediction accuracy in traditional prediction. On this basis, combined with the demand change rate Dr(t) and the equipment warehouse age Ts(t), the forecast value is further adjusted progressively to calculate the comprehensive adjusted demand forecast value Dadj(t). Finally, by adding environmental influencing factors such as warehouse temperature Hw(t) and warehouse humidity Hs(t), the final demand forecast value Dh(t) is further calculated progressively. This multi-layer progressive algorithm model significantly improves the flexibility and accuracy of the prediction. The progressive algorithm not only considers static historical data, but also dynamically introduces multi-dimensional real-time data to ensure the accuracy and timeliness of demand forecasting. Compared with the traditional static prediction method, this module gradually optimizes the prediction results through a progressive algorithm, avoiding the problem of poor adaptation of a single model to complex environments in traditional methods, and greatly improving the accuracy and adaptability of predictions. By considering multi-dimensional parameters such as equipment age, temperature and humidity, this solution can better reflect the actual use environment, reduce prediction errors, and thus achieve a more accurate judgment of future demand.

[0114] Example 4

[0115] This embodiment is explained in Example 3, please refer to Figure 1 , specifically: S3 includes S31, S32 and S33;

[0116] S31. Establish a multi-level warning mode according to the current inventory quantity S(t) and the inventory threshold L to obtain the warning level Pw(t). The inventory threshold L includes the inventory upper limit threshold Lmax and the inventory lower limit threshold Lmin.

[0117] The warning level Pw(t) is obtained by the multi-level warning mode;

[0118]

[0119] In the formula, 0 represents the first-level inventory warning, 1 represents the second-level inventory warning, and 2 represents the third-level inventory warning;

[0120] Level 1 inventory warning > Level 2 inventory warning > Level 3 inventory warning;

[0121] Level 1 inventory warning means that the inventory is in the normal range. At this time, it is considered that the inventory is sufficient to meet future demand and no immediate replenishment or adjustment measures are required;

[0122] Level 2 inventory warning means that the inventory level is close to the minimum safety stock threshold. Although the inventory level is still within an acceptable range, it is close to the warning state. It is necessary to pay attention to the inventory consumption and may need to take replenishment actions in the near future.

[0123] The third-level inventory warning means that the inventory level is lower than the minimum safety threshold. At this time, the inventory has entered an emergency state. The inventory level is insufficient to meet demand, which may cause a supply chain disruption. Immediate replenishment measures must be taken to avoid out-of-stock problems.

[0124] S32, calculate the inventory difference by using the predicted final demand forecast value Dh(t) and the current inventory quantity S(t) to obtain the supply-demand difference value E1;

[0125] The supply and demand difference value E1 is calculated and output by the following difference calculation formula;

[0126] E1=S(t)-Dh(t);

[0127] S33. Based on the output result of the obtained supply-demand difference value E1, a preliminary comparative evaluation is performed to analyze the inventory status of the current measurement assets. The specific evaluation contents are as follows;

[0128] When the supply-demand difference value E1>0, it means that the inventory of the predicted metering asset equipment is sufficient, and there is no need to send an early warning;

[0129] When the supply-demand difference value E1≤0, it means that the predicted inventory of metering asset equipment is insufficient, and secondary optimization is performed at this time.

[0130] In this embodiment, the method constructs a multi-level early warning mode based on the current inventory S(t) and the inventory threshold L, which can automatically determine whether the inventory is in a normal, early warning or emergency state, and dynamically generate the early warning level Pw(t). This multi-level early warning mechanism can promptly remind managers to take corresponding actions when the inventory is close to or below the safety threshold, avoiding the problem of delayed response in inventory management. By calculating the supply and demand difference E1 between the current inventory S(t) and the final demand forecast value Dh(t), it is possible to accurately determine whether the current inventory is sufficient to meet future demand, thereby providing a basis for subsequent decision-making. Through the evaluation of the supply and demand difference value E1, the current inventory status can be quickly analyzed. When the inventory is insufficient, the system will automatically trigger further optimization and adjustment steps to ensure the rationality and adequacy of the inventory. Compared with the traditional static inventory monitoring method, this module realizes the foresight and flexibility of inventory management through a dynamic multi-level early warning mechanism and supply and demand difference analysis, avoiding the phenomenon of insufficient or backlog inventory. In traditional methods, inventory early warnings often lack hierarchy and precision, while through multi-level early warnings, the system can accurately identify different stages of inventory fluctuations and issue early warning signals in advance. By calculating the supply-demand difference value E1, the system ensures seamless integration of inventory and demand forecasts, greatly improving the response speed and accuracy of inventory management, reducing operational risks, and improving the efficiency and reliability of supply chain management.

[0131] Example 5

[0132] This embodiment is explained in Example 4. Please refer to Figure 1 , specifically: S4 includes S41 and S42;

[0133] S41. When the preliminary comparative evaluation result shows that the inventory is insufficient or abnormal, the optimized demand forecast value Dopt(t) is further calculated based on the equipment status of the metering assets and the final demand forecast value Dh(t) to optimize the metering assets;

[0134] The optimized demand forecast value Dopt(t) is calculated and output by the following algorithm formula;

[0135]

[0136] In the formula, κ represents the adjustment coefficient of failure rate to forecast demand, λ represents the adjustment coefficient of inventory to forecast demand, Fmax represents the upper limit of failure rate, and Smax represents the upper limit of inventory;

[0137] S42, performing a secondary difference calculation based on the obtained optimized demand forecast value Dopt(t) and the current inventory S(t) of the metered asset to obtain a second supply-demand difference value E2;

[0138] The second supply-demand difference value E2 is calculated by the following difference calculation formula:

[0139] E2 = S(t) - Dopt(t).

[0140] In this embodiment, when the method finds that the inventory is insufficient or abnormal through preliminary comparative evaluation, it further calculates the optimized demand forecast value Dopt(t) by combining the actual condition of the metering asset equipment and the final demand forecast value Dh(t). This optimization process takes into account the adjustment coefficient of the equipment failure rate and the inventory, ensuring that the predicted demand value can be dynamically adjusted when the equipment failure rate is high or the inventory is insufficient, thereby improving the accuracy of the prediction. Then, by performing a secondary difference calculation between the optimized demand forecast value Dopt(t) and the current inventory S(t), a second supply and demand difference value E2 is obtained to further evaluate whether the current inventory is sufficient to meet the optimized demand. Through this secondary evaluation, the system can more accurately determine whether it is necessary to replenish stocks immediately or make other inventory adjustments to ensure that inventory and demand are accurately matched. This method step makes demand forecasting more flexible and accurate by introducing an adjustment mechanism for equipment failure rate and inventory, avoiding the problem that the traditional method fails to fully consider the impact of equipment status on demand. The secondary supply and demand difference analysis further improves the accuracy of inventory management, ensuring that the optimized forecast can truly reflect actual demand. Overall, this improvement effectively reduces the risk of insufficient or excessive inventory, ensures the continuity and stability of the supply of measurement assets, and improves the intelligence and automation level of the system.

[0141] Example 6

[0142] This embodiment is explained in Example 5. Please refer to Figure 1 , specifically: S5 includes S51;

[0143] S51. Based on the obtained second supply-demand difference value E2 output result, and in combination with the multi-level early warning mode, a secondary comparative evaluation is performed to analyze the inventory status of the final measurement asset equipment. The specific evaluation contents are as follows;

[0144] When the second supply-demand difference value E2>0, it means that the inventory of the measurement asset equipment is predicted to be sufficient. At this time, a first-level inventory warning is sent through the measurement asset management and control platform;

[0145] When the second supply-demand difference value E2≤0, it means that the inventory of the predicted measurement asset equipment is insufficient. At this time, a secondary inventory warning is sent through the measurement asset management and control platform;

[0146] When the second supply-demand difference value E2≤2, it indicates that the predicted inventory of metering asset equipment is abnormal. At this time, a third-level inventory warning is sent through the metering asset management and control platform.

[0147] In this embodiment, the method outputs the evaluation result based on the second supply-demand difference value E2, and conducts a secondary evaluation of the inventory of the metering asset in combination with the multi-level early warning mode. By further analyzing the inventory situation, when E2>0, the system determines that the inventory is sufficient and sends a first-level inventory warning through the metering asset management and control platform, indicating that no replenishment is required. When E2≤0, the system believes that the inventory is insufficient and sends a second-level inventory warning, prompting that attention should be paid to inventory consumption and replenishment measures should be prepared. When E2≤2, it indicates that the inventory is abnormal, and the system will send a third-level warning, indicating that the inventory is in an emergency state and replenishment actions need to be taken immediately. Through the secondary comparative evaluation, the system can more accurately analyze the inventory status, ensure timely response, and avoid inventory breaks or backlogs. Compared with the traditional inventory management method, the improvement of this step lies in the combination of multi-level warning and secondary evaluation, which further refines the decision-making mechanism of inventory management. Traditional methods often rely only on a single warning or lack of in-depth evaluation, resulting in delayed or overreaction of inventory response. Through the calculation of the second supply-demand difference value E2, the system can more flexibly and accurately identify inventory changes, and automatically generate corresponding management measures for different warning levels.

[0148] Example 7

[0149] See also Figure 1 and Figure 2 , an algorithm-based intelligent demand forecasting system for metering assets, including a demand collection module, a demand forecasting algorithm module, a multi-level early warning and preliminary analysis module, a secondary forecasting optimization analysis module and a multi-level early warning combination module;

[0150] The demand collection module builds a metering asset management and control platform and sets up an API application program interface to connect with the relevant systems of metering assets, collects metering asset demand data in relevant systems in real time, pre-processes the metering asset demand data, integrates it into a demand data set, and then sets up a time series database to store the demand data set;

[0151] The demand forecasting algorithm module constructs a basic demand forecasting algorithm model, performs preliminary calculations and outputs the basic demand forecast value Db(t), and uses a progressive algorithm based on the basic demand forecast value Db(t) to sequentially output the comprehensive adjustment demand forecast value Dadj(t) and the final demand forecast value Dh(t);

[0152] The multi-level warning and preliminary analysis module establishes a multi-level warning mode based on the current inventory of the metering asset S(t) and the inventory threshold L, obtains the warning level Pw(t), and then calculates the difference between the current inventory of the metering asset S(t) and the obtained final demand forecast value Dh(t), obtains the supply-demand difference value E1, and based on the output of the supply-demand difference value, conducts a preliminary evaluation and analysis of the metering assets and inventory in the future forecast stage;

[0153] The secondary forecast optimization analysis module calculates and outputs the optimized demand forecast value Dopt(t) based on the abnormal result of the preliminary evaluation, and calculates the difference between the current inventory of the metering asset S(t) and the optimized demand forecast value Dopt(t) to obtain the second supply and demand difference value E2;

[0154] The multi-level early warning combination module performs a secondary evaluation based on the second supply-demand difference value E2 to optimize the demand for metering assets, and generates relevant information prompts through the metering asset management and control platform.

[0155] While the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that many changes, modifications, substitutions and variations can be made to the embodiments without departing from the principles and spirit of the invention.

Claims

1. An algorithm-based intelligent demand forecasting method for metering assets, characterized in that: The following steps are involved: S1. Build a metering asset management and control platform and set up an API application program interface to connect with the relevant systems of metering assets, collect metering asset demand data in relevant systems in real time, pre-process the metering asset demand data, integrate it into a demand data set, and then set up a time series database to store the demand data set; S2. Construct a basic demand forecasting algorithm model, perform preliminary calculations to output a basic demand forecast value Db(t), and use a progressive algorithm based on the basic demand forecast value Db(t) to sequentially output a comprehensive adjustment demand forecast value Dadj(t) and a final demand forecast value Dh(t); S3. A multi-level early warning mode is established based on the current inventory S(t) of the measurement asset and the inventory threshold L to obtain the early warning level Pw(t). The difference between the current inventory S(t) of the measurement asset and the obtained final demand forecast value Dh(t) is calculated to obtain the supply-demand difference value E1. Based on the output of the supply-demand difference value, a preliminary evaluation and analysis of the measurement assets and inventory in the future forecast stage is performed. S4. When the preliminary evaluation result is abnormal, the optimized demand forecast value Dopt(t) is calculated and output based on the current use of the power grid equipment, and the difference between the current inventory of the metering asset S(t) and the optimized demand forecast value Dopt(t) is calculated to obtain the second supply and demand difference value E2; S5. Based on the second supply-demand difference value E2, a second evaluation is performed on the demand for measurement assets after optimization, and relevant information prompts are generated through the measurement asset management and control platform.

2. According to claim 1, a method for intelligent demand forecasting of metering assets based on an algorithm is characterized in that: Said S1 includes S11, S12 and S13; S11. Build a metering asset management and control platform, set up an API application program interface to connect with the relevant systems of metering assets, and set up to collect metering asset demand data from relevant systems every hour; The related systems include storage systems, power grid systems and meteorological systems; The metering asset demand data includes inventory S, demand change rate Dr, metering asset equipment storage age Ts, warehouse temperature Hw, warehouse humidity Hs, failure rate Fr, power grid fluctuation rate Ve and environmental disaster index Wd; The inventory quantity S and the inventory age Ts of the measuring asset equipment are collected and obtained through the warehousing system; The demand change rate Dr is calculated using the historical outbound data provided by the warehousing system and the time series analysis algorithm to output the demand change rate Dr of the metering asset; The warehouse temperature Hw is obtained by collecting the temperature parameters in the warehouse through the temperature sensor integrated in the storage system in real time; The warehouse humidity Hs is obtained by collecting humidity parameters in the warehouse through the humidity sensor integrated in the storage system in real time; The failure rate Fr is calculated and obtained by collecting historical usage and maintenance records of the power grid system; The grid fluctuation rate Ve is measured by collecting the power load P(t) at time t and the reference load value Pavg by the grid system; The environmental disaster index Wd is collected through a meteorological system.

3. The algorithm-based intelligent demand forecasting method for metering assets according to claim 2 is characterized in that: S12. Preprocessing the collected metrology asset demand data in the metrology asset management and control platform, wherein the preprocessing method includes marking timestamps and standardization processing, and after the preprocessing, integrating and integrating to obtain the demand data set; The marking timestamp is used to mark each parameter in the collected metering asset demand data according to the collection time, and obtain the inventory S(t), demand change rate Dr(t), metering asset equipment age Ts(t), warehouse temperature Hw(t), warehouse humidity Hs(t), failure rate Fr(t), power grid fluctuation rate Ve(t) and environmental disaster index Wd(t) at time t respectively; S13. By constructing a time series database and setting a write port, a write port, a real-time data table and a historical data table, the demand data set obtained after preprocessing is stored in the real-time data table through the write port, and the demand data set stored last time is automatically transferred to the historical data table and stored in chronological order.

4. The algorithm-based intelligent demand forecasting method for metering assets according to claim 3 is characterized in that: The S2 includes S21, S22 and S23; S21. Extract the demand data set in the real-time data table through the write port, input it into the basic demand forecasting algorithm model, calculate and output the basic demand forecast value Db(t), and predict the demand for the measurement asset at the next t time.

5. The algorithm-based intelligent demand forecasting method for metering assets according to claim 4 is characterized in that: S22. Based on the obtained basic demand forecast value Db(t), a progressive algorithm is used to calculate and output the comprehensive adjusted demand forecast value Dadj(t) in combination with the demand change rate Dr(t) and the inventory age of the metering asset equipment Ts(t). Then, a progressive calculation is performed based on the comprehensive adjusted demand forecast value Dadj(t) in combination with the warehouse temperature Hw(t) and the warehouse humidity Hs(t) to obtain the final demand forecast value Dh(t). The comprehensive adjustment demand forecast value Dadj(t) is further calculated and obtained through the basic demand forecast value Db(t); The final demand forecast value Dh(t) is further calculated and obtained by comprehensively adjusting the demand forecast value Dadj(t); The comprehensive adjustment demand forecast value Dadj(t) and the final demand forecast value Dh(t) are calculated and obtained by the following progressive algorithm: In the formula, represents the differential variable of the demand change rate, d represents the differential function, dt represents the differential variable of time t, η represents the temperature influence adjustment coefficient, μ represents the humidity influence adjustment coefficient, Hwopt represents the ideal temperature condition, Hsopt represents the ideal humidity condition, δ represents the adjustment coefficient of the demand change rate, ε represents the adjustment coefficient of the inventory age of the metering asset equipment, and ζ represents the adjustment coefficient of the failure rate.

6. The algorithm-based intelligent demand forecasting method for metering assets according to claim 1 is characterized in that: The S3 includes S31, S32 and S33; S31. Establish a multi-level warning mode according to the current inventory S(t) and the inventory threshold L to obtain the warning level Pw(t), wherein the inventory threshold L includes an upper inventory threshold Lmax and a lower inventory threshold Lmin; The warning level Pw(t) is obtained through the multi-level warning mode; In the formula, 0 represents the first-level inventory warning, 1 represents the second-level inventory warning, and 2 represents the third-level inventory warning; Level 1 inventory warning > Level 2 inventory warning > Level 3 inventory warning.

7. The algorithm-based intelligent demand forecasting method for metering assets according to claim 6 is characterized in that: S32, calculate the inventory difference by using the predicted final demand forecast value Dh(t) and the current inventory quantity S(t) to obtain the supply-demand difference value E1; S33. Based on the output result of the obtained supply-demand difference value E1, a preliminary comparative evaluation is performed to analyze the inventory status of the current measurement assets. The specific evaluation contents are as follows; When the supply-demand difference value E1>0, it means that the inventory of the predicted metering asset equipment is sufficient, and there is no need to send an early warning; When the supply-demand difference value E1≤0, it means that the predicted inventory of metering asset equipment is insufficient, and secondary optimization is performed at this time.

8. The algorithm-based intelligent demand forecasting method for metering assets according to claim 6 is characterized in that: The S4 includes S41 and S42; S41. When the preliminary comparative evaluation result shows that the inventory is insufficient or abnormal, the optimized demand forecast value Dopt(t) is further calculated based on the equipment status of the metering assets and the final demand forecast value Dh(t) to optimize the metering assets; The optimized demand forecast value Dopt(t) is calculated and outputted by the following algorithm formula; In the formula, κ represents the adjustment coefficient of failure rate to forecast demand, λ represents the adjustment coefficient of inventory to forecast demand, Fmax represents the upper limit of failure rate, and Smax represents the upper limit of inventory; S42. Perform a secondary difference calculation based on the obtained optimized demand forecast value Dopt(t) and the current inventory S(t) of the metered asset to obtain a second supply-demand difference value E2.

9. The algorithm-based intelligent demand forecasting method for metering assets according to claim 8 is characterized in that: The S5 includes S51; S51. Based on the obtained second supply-demand difference value E2 output result, and in combination with the multi-level early warning mode, a secondary comparative evaluation is performed to analyze the inventory status of the final measurement asset equipment. The specific evaluation contents are as follows; When the second supply-demand difference value E2>0, it means that the inventory of the measurement asset equipment is predicted to be sufficient. At this time, a first-level inventory warning is sent through the measurement asset management and control platform; When the second supply-demand difference value E2≤0, it means that the inventory of the predicted measurement asset equipment is insufficient. At this time, a secondary inventory warning is sent through the measurement asset management and control platform; When the second supply-demand difference value E2≤2, it indicates that the predicted inventory of metering asset equipment is abnormal. At this time, a third-level inventory warning is sent through the metering asset management and control platform.

10. An algorithm-based intelligent demand forecasting system for metering assets, applied to an algorithm-based intelligent demand forecasting method for metering assets according to any one of claims 1 to 9, characterized in that: It includes demand collection module, demand forecasting algorithm module, multi-level early warning and preliminary analysis module, secondary forecasting optimization analysis module and multi-level early warning combination module; The demand collection module builds a metering asset management and control platform and sets up an API application program interface to connect with the relevant systems of metering assets, collects metering asset demand data in relevant systems in real time, pre-processes the metering asset demand data, integrates it into a demand data set, and then sets up a time series database to store the demand data set; The demand forecasting algorithm module constructs a basic demand forecasting algorithm model, performs preliminary calculations to output a basic demand forecast value Db(t), and uses a progressive algorithm based on the basic demand forecast value Db(t) to sequentially output a comprehensive adjustment demand forecast value Dadj(t) and a final demand forecast value Dh(t); The multi-level warning and preliminary analysis module establishes a multi-level warning mode according to the current inventory S(t) of the metering asset and the inventory threshold L, obtains the warning level Pw(t), and then calculates the difference between the current inventory S(t) of the metering asset and the obtained final demand forecast value Dh(t), obtains the supply-demand difference value E1, and based on the output of the supply-demand difference value, performs a preliminary evaluation and analysis of the metering assets and inventory conditions in the future forecast stage; The secondary forecast optimization analysis module calculates and outputs the optimized demand forecast value Dopt(t) based on the abnormal result of the preliminary evaluation, and calculates the difference between the current inventory S(t) of the metering asset and the optimized demand forecast value Dopt(t) to obtain the second supply-demand difference value E2; The multi-level early warning combination module performs a secondary evaluation on the optimized metrological asset demand based on the second supply-demand difference value E2, and generates relevant information prompts through the metrological asset management and control platform.

Images