Enterprise production index monitoring method based on electric power data

Abstract

The invention provides an enterprise production index monitoring method based on electric power data, and relates to the technical field of enterprise production index monitoring. The method comprises the following steps: acquiring real-time power data, water consumption, heat / gas energy consumption data and equipment sensor data of an enterprise, including vibration, temperature and pressure sensor data; calculating a production trend index and a product energy efficiency index by using the dynamic weight, and generating a comprehensive energy efficiency index CEI in combination with the equipment health coefficient; and when the CEI is lower than a threshold value, fault diagnosis is carried out based on the Bayesian network and early warning is triggered. And meanwhile, the electric power data is stored through a block chain technology, and data credibility verification and exception recheck are realized in combination with financial data cross validation. The system realizes multi-dimensional production state monitoring, has the advantages of dynamic adaptation to production fluctuation, intelligent fault diagnosis, data credibility guarantee and the like, can effectively improve enterprise production management efficiency, reduces energy consumption and fault loss, and is suitable for digital transformation requirements of high-energy-consumption industries such as chemical industry and manufacturing industry.

Description

Technical Field

[0001] The present invention relates to the technical field of enterprise production indicator monitoring, and specifically to an enterprise production indicator monitoring method based on power data, which is used for production status monitoring and optimization in high-energy-consuming industries such as chemical industry, manufacturing, and steel. Background Art

[0002] Accurately monitoring production metrics is crucial for ensuring efficiency, controlling costs, and ensuring safety during production. Traditional production metric monitoring methods often rely on a single data source or simple statistical analysis, resulting in limitations such as a single monitoring dimension and a lack of dynamic adaptability. For example, calculating production metrics solely based on electricity data fails to fully reflect equipment health and other energy consumption patterns. Fixed-weighted metric calculations are also difficult to adapt to cyclical fluctuations and environmental changes in production.

[0003] Furthermore, existing monitoring systems lack data credibility verification, making data susceptible to tampering and collection errors, leading to inaccurate monitoring results. They also lack intelligent means for diagnosing and handling abnormalities, making it impossible to quickly locate the root cause of faults, impacting the continuity and stability of enterprise production. Therefore, a data processing method and system that can integrate multi-source data, dynamically adjust calculation parameters, and implement intelligent diagnosis is urgently needed to improve the accuracy and reliability of enterprise production indicator monitoring. Summary of the Invention

[0004] The purpose of the present invention is to provide a method and system for monitoring enterprise production indicators based on power data. By collecting multi-source energy consumption data and equipment sensor data, combined with dynamic weight calculation and intelligent diagnosis algorithms, comprehensive and accurate monitoring of enterprise production indicators can be achieved, and abnormal situations in the production process can be discovered in a timely manner, providing a scientific basis for enterprise production management.

[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is: The enterprise production indicator monitoring method based on power data specifically includes the following steps: Step 1: Collect enterprise power data in real time; Step 2: Dynamically assign weights based on historical electricity consumption fluctuation intensity and calculate the production trend index TI; Step 3: Calculate the product energy efficiency index (EEI) based on industry benchmarks; Step 4: TI, EEI, and the equipment health factor Hscore are integrated to generate the comprehensive energy efficiency index CEI, where Hscore is dynamically calculated based on the deviation value of the equipment sensor; Step 5: When the CEI is lower than the threshold, fault diagnosis is performed through the Bayesian network and an early warning is triggered.

[0006] Further, in step 1, collect the real-time electricity data of the enterprise, including the electricity consumption CU this month, the electricity consumption PU last month, the electricity consumption LYC in the same month last year, and the electricity consumption LYN in the next month last year; In step 2, it specifically includes the following steps to calculate the production trend index TI; Calculate the expected month-on-month ratio: EC = (LYN / LYC) × 100% Calculate the current month-on-month ratio: CC = (CU / PU) × 100% Dynamically allocate weights W3 and W4, and satisfy W3 + W4 = 100%; Generate TI = EC × W3 + CC × W4; In step 3, calculate the product energy efficiency index EEI specifically as follows; Obtain the unit energy consumption of the industry standard product: SPE = total monthly electricity consumption of all manufacturers in the same industry ÷ total output; Calculate the actual unit energy consumption of the product: UE = CU ÷ monthly output Q; Generate EEI = (SPE / UE) × 100%; In step 4, generate the production monitoring index: PMI = TI × W1 + EEI × W2, where Wl + W2 = 100%; Further, the assignment method of the dynamic weight W3 in step 2 is as follows. Based on the electricity consumption data in the past 12 months, calculate its variance VAR as the fluctuation intensity index VI; VI = VAR(E1, E2, …, E12); Where, Ei is the electricity consumption in the i-th month, VI is the fluctuation intensity, and VAR is the variance calculation function; Determine the dynamic weight W3 according to the fluctuation intensity (VI): If VI ≥ high threshold V1, then W3 = 70%, paying more attention to the expected month-on-month ratio; If VI ≤ low threshold V2, then W3 = 30%, paying more attention to the current month-on-month ratio; If low threshold V2 < VI < high threshold V1, then calculate by linear interpolation: ; According to W3 + W4 = 100%, calculate W4 = 100% W3.

[0007] Further, the calculation method of the unit energy consumption of the standard product is specifically as follows: S101: Data collection. Obtain the monthly data of at least 10 representative enterprises in the same industry through the cloud server, including: The monthly total output (Q1, Q2, …, Qn) of each enterprise; Monthly total electricity consumption (E1, E2, …, En); S102: Calculate industry benchmarks: Calculate the total electricity consumption of the industry: ∑E=E1+E2+…+En; Calculate the total industry output: ∑Q=Q1+Q2+…+Qn; Calculate standard product specific energy (SPE): ; S103: Data update: Automatically update SPE values every month to ensure the timeliness of the benchmark.

[0008] Furthermore, step 1 also includes collecting water consumption, heat / gas energy consumption data and equipment sensor data, wherein the equipment sensors include vibration sensors, temperature sensors and pressure sensors; On the basis of the original power data, the following data are collected: Water consumption data: total water consumption for the month (WU), water consumption per unit product (WUP) Heat / gas energy consumption data: total heat / gas volume (HU) for the month, heat / gas consumption per unit product (HUP) Device sensor data: Vibration sensor: real-time vibration value (V), reference vibration value (Vstd); Temperature sensor: real-time temperature value (T), reference temperature value (Tstd); Pressure sensor: real-time pressure value (P), reference pressure value (Pstd); Step 4 is replaced by generating the comprehensive energy efficiency index CEI: ; in, is the dynamic weight, ; Estd, Wstd, and Tstd are pre-stored industry energy consumption benchmarks, and Eactual, Wactual, and Tactual are actual energy consumption values.

[0009] Furthermore, the calculation process of the equipment health coefficient Hscore is as follows: ; ΔV=(Vactual-Vstd) / Vstd; ΔT=(Tactual-Tstd) / Tstd; ΔP=(Pactual-Pstd) / Pstd; Among them: Vactual, Tactual, Pactual are actual measured values, Vstd, Tstd, Pstd are standard values.

[0010] Furthermore, dynamic weight Adjustment methods include: Calculate the energy consumption fluctuation coefficients Evar, Wvar, Tvar, and Hvar, By correcting the function Generate scaling factors; according to Calculate intermediate weights; The final weights α, β, γ and ; The details are as follows: ; in: The preset weights are for the industries, and Evar, Wvar, Tvar, and Hvar are energy consumption fluctuation coefficients.

[0011] Furthermore, when the CEI is lower than the threshold, fault diagnosis is performed. The specific steps are as follows: P(F|E)∝P(E|F)×P(F); Fdetected=argmax[P(F|E)]; Where: F is the fault type, E is the mutation feature of the evidence monitoring data, P(F|E) is the posterior probability of fault type F under evidence E, argmax[P(F|E)] is the fault type with the largest posterior probability as the diagnosis result Fdetected, P(E|F) is the conditional probability of evidence E when fault F occurs, and P(F) is the prior probability of the fault.

[0012] Furthermore, the conditional probability includes: P(CF|EP>20%)=0.8; P(CF|TD>5℃)=0.9; P(PF|VO>50%)=0.85; Among them: CF is cooling system failure, EP is power surge, TD is water temperature deviation, PF is pump failure, and VO is vibration exceeding the standard.

[0013] Furthermore, the data credibility verification step is also included: ; Vresult=Compare(Fdata,BCdata); Among them: DFP is the production digital fingerprint, Edata is the power data, Fdata is the financial data, BCdata is the blockchain data, Vresult is the difference result of comparing the financial data Fdata and the blockchain data BCdata; the Compare function calculates the relative difference between the key fields of the financial data Fdata and the blockchain data BCdata.

[0014] It also includes the following exception review steps: RI=|(ΔQ / ΔE)-1|; When RI>0.3, the manual review process is triggered; Among them: RI is the review index, ΔQ is the change in output, and ΔE is the change in energy consumption.

[0015] Furthermore, the present invention also discloses an enterprise production index monitoring system, which includes: Data acquisition module: used to collect power data, water consumption data, heat / gas energy consumption data and equipment sensor data; Data processing module: realizes the calculation of production trend index, product energy efficiency index, comprehensive energy efficiency index and other indicators; Intelligent diagnosis module: Fault diagnosis and early warning based on Bayesian network; Data storage and verification module: stores data in the blockchain and verifies its credibility, while also performing exception review.

[0016] Compared with the prior art, the present invention has the following beneficial effects: The present invention integrates electricity, water, heat / gas energy consumption and equipment sensor data to comprehensively reflect the production status of the enterprise, avoid the limitations of single data, adapt to the periodic fluctuations and environmental changes in the enterprise production process through dynamic weight adjustment, and improve the accuracy of indicator calculation. At the same time, the fault diagnosis algorithm based on the Bayesian network can quickly locate the root cause of the fault, with a fault diagnosis accuracy rate of 90%, and early warning to avoid production losses; more importantly, blockchain technology ensures that data cannot be tampered with, and the tax data comparison and cross-validation mechanism ensures that the monitoring results are true and reliable. The present invention provides enterprises with a scientific production management basis, achieves energy efficiency optimization, reduces maintenance costs, improves spare parts inventory turnover and other goals, and helps enterprises with digital transformation. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following is a brief introduction to the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1 This is a block diagram of the overall principle of the present invention. DETAILED DESCRIPTION

[0019] Hereinafter, only certain exemplary embodiments are briefly described. As will be appreciated by those skilled in the art, the described embodiments may be modified in various ways without departing from the spirit or scope of the embodiments of the present invention. Therefore, the drawings and description are to be regarded as illustrative in nature and not restrictive. Embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0020] Example 1: See Figure 1 , a method for monitoring enterprise production indicators based on power data, comprising: Step S1: Real-time collection of enterprise production data, including power data, water consumption data, heat / gas consumption data, and equipment operation status data; Step S2: Dynamically calculate the production trend index, assign weights based on the fluctuation characteristics of historical energy consumption data, and generate a trend index by combining the expected month-on-month and current month-on-month data; Step S3: Calculate the product energy efficiency index and generate an energy efficiency evaluation index by comparing the actual unit energy consumption with the industry benchmark energy consumption; Step S4: Generate a comprehensive energy efficiency index CEI by integrating the production trend index, product energy efficiency index and equipment health coefficient, wherein the equipment health coefficient is dynamically calculated based on the equipment operation status deviation value; Step S5: Fault diagnosis and early warning. When the CEI is lower than the preset threshold, fault probability inference is performed based on the mutation characteristics of multi-source data and an early warning is triggered. Step S6: Data credibility verification, storing key data through blockchain technology and cross-verifying with external trusted data sources.

[0021] Furthermore, the dynamic weight allocation in step S2 includes calculating the fluctuation intensity index of the historical energy consumption series; and dynamically adjusting the weight ratio of the expected year-on-year growth and the current year-on-year growth according to the fluctuation intensity range, wherein the greater the fluctuation intensity, the higher the weight of the expected year-on-year growth.

[0022] Furthermore, the industry benchmark energy consumption in step S3 is obtained by aggregating the monthly total energy consumption and total output data of multiple enterprises in the same industry, calculating the industry average unit product energy consumption as a benchmark value, and regularly updating the benchmark value.

[0023] The equipment health coefficient is calculated by obtaining the deviation between the real-time measurement value of the equipment sensor and the standard value, and fusing the multi-dimensional deviations based on preset weights to obtain a score value representing the health of the equipment.

[0024] Furthermore, the failure probability reasoning in step S5 is implemented using a Bayesian network, including: Construct a probability table of association between fault types and data mutation characteristics; Calculate the maximum a posteriori probability fault type based on real-time data mutation evidence.

[0025] The cross validation in step S6 includes: Compare energy consumption data stored on the blockchain with financial reporting data; When the difference exceeds the threshold, manual review is triggered.

[0026] Furthermore, in actual implementation, the triggering condition for manual review also includes calculating the deviation of the ratio of the output change to the energy consumption change, and starting the review process when the deviation exceeds the preset tolerance range.

[0027] Example 2: This example discloses a method for monitoring enterprise production indicators based on power data, which specifically includes the following steps: Step 1: Collect enterprise power data in real time; Step 2: Dynamically assign weights based on historical electricity consumption fluctuation intensity and calculate the production trend index TI; Step 3: Calculate the product energy efficiency index (EEI) based on industry benchmarks; Step 4: TI, EEI, and the equipment health factor Hscore are integrated to generate the comprehensive energy efficiency index CEI, where Hscore is dynamically calculated based on the deviation value of the equipment sensor; Step 5: When the CEI is lower than the threshold, fault diagnosis is performed through the Bayesian network and an early warning is triggered.

[0028] Furthermore, in step 1, the real-time electricity data of the enterprise is collected, including the electricity consumption of this month CU, the electricity consumption of last month PU, the electricity consumption of this month last year LYC, and the electricity consumption of the next month last year LYN; In step 2, the specific steps include: calculating the production trend index TI; Calculate expected month-on-month growth: EC = (LYN / LYC) × 100% Calculate the current situation year-on-year: CC = (CU / PU) × 100% Dynamically allocate weights W3 and W4, and satisfy W3+W4=100%; Generate TI = EC × W3 + CC × W4; In step 3, the product energy efficiency index EEI is calculated as follows; Obtain industry standard product unit energy consumption: SPE = total monthly electricity consumption of all manufacturers in the same industry ÷ total output; Calculate the actual unit energy consumption of the product: UE = CU ÷ monthly output Q; Generate EEI = (SPE / UE) × 100%; In step 4, the production monitoring index is generated: PMI = TI × W1 + EEI × W2, where W1 + W2 = 100%; Further, the assignment method of the dynamic weight W3 in step 2 is as follows. Based on the electricity consumption data of the past 12 months, calculate its variance VAR as the fluctuation intensity index VI; VI = VAR(E1, E2, …, E12); where Ei is the electricity consumption of the i-th month, VI is the fluctuation intensity, and VAR is the variance calculation function; Determine the dynamic weight W3 according to the fluctuation intensity (VI): If VI ≥ high threshold V1, then W3 = 70%, paying more attention to the expected month-on-month; If VI ≤ low threshold V2, then W3 = 30%, paying more attention to the current month-on-month; If low threshold V2 < VI < high threshold V1, calculate by linear interpolation: ; According to W3 + W4 = 100%, calculate W4 = 100% W3.

[0029] Further, the calculation method of the unit energy consumption of the standard product is specifically as follows: S101: Data collection. Obtain the monthly data of at least 10 representative enterprises in the same industry through the cloud server, including: The monthly total output (Q1, Q2, …, Qn) of each enterprise; The monthly total electricity consumption (E1, E2, …, En); S102: Calculate the industry benchmark: Calculate the total industry electricity consumption: ∑E = E1 + E2 + … + En; Calculate the total industry output: ∑Q = Q1 + Q2 + … + Qn; Calculate the unit energy consumption of the standard product (SPE): ; S103: Data update. Automatically update the SPE value every month to ensure the timeliness of the benchmark.

[0030] Further, step 1 also includes collecting water consumption, heat / gas energy consumption data and equipment sensor data. The equipment sensors include vibration sensors, temperature sensors and pressure sensors; On the basis of the original power data, add the collection of the following data: Water consumption data: the total monthly water consumption (WU), the water consumption per unit product (WUP) Heat / gas energy consumption data: the total monthly heat / gas volume (HU), the heat / gas consumption per unit product (HUP) Equipment sensor data: Vibration sensor: real-time vibration value (V), reference vibration value (Vstd); Temperature sensor: real-time temperature value (T), reference temperature value (Tstd); Pressure sensor: real-time pressure value (P), reference pressure value (Pstd); Step 4 is replaced by generating the comprehensive energy efficiency index CEI: ;in, is the dynamic weight, ; Estd, Wstd, and Tstd are pre-stored industry energy consumption benchmarks, and Eactual, Wactual, and Tactual are actual energy consumption values.

[0031] Furthermore, the calculation process of the equipment health coefficient Hscore is as follows: ; ΔV=(Vactual-Vstd) / Vstd; ΔT=(Tactual-Tstd) / Tstd; ΔP=(Pactual-Pstd) / Pstd; Among them: Vactual, Tactual, Pactual are actual measured values, Vstd, Tstd, Pstd are standard values.

[0032] Furthermore, dynamic weight Adjustment methods include: Calculate the energy consumption fluctuation coefficients Evar, Wvar, Tvar, and Hvar, By correcting the function Generate scaling factors; according to Calculate intermediate weights; The final weights α, β, γ and ; The details are as follows: ; in: The preset weights are for the industries, and Evar, Wvar, Tvar, and Hvar are energy consumption fluctuation coefficients.

[0033] Furthermore, when the CEI is lower than the threshold, fault diagnosis is performed. The specific steps are as follows: P(F|E)∝P(E|F)×P(F); Fdetected=argmax[P(F|E)]; Where: F is the fault type, E is the evidence, and monitoring data mutation characteristics.

[0034] Furthermore, the conditional probability includes: P(CF|EP>20%)=0.8; P(CF|TD>5℃)=0.9; P(PF|VO>50%)=0.85; Among them: CF is cooling system failure, EP is power surge, TD is water temperature deviation, PF is pump failure, and VO is vibration exceeding the standard.

[0035] Furthermore, the data credibility verification step is also included: ; Vresult=Compare(Fdata,BCdata); Among them: DFP is the production digital fingerprint, Refers to the real-time electricity data collected by the enterprise for each month; Fdata is financial data, BCdata is blockchain data, and Vresult is the difference between the financial data Fdata and the blockchain data BCdata; the Compare function calculates the relative difference between the key fields of the financial data Fdata and the blockchain data BCdata; A digital fingerprint (DFP) is generated through a hash algorithm for blockchain evidence storage.

[0036] It also includes the following exception review steps: RI=|(ΔQ / ΔE)-1|; When RI>0.3, the manual review process is triggered; Among them: RI is the review index, ΔQ is the change in output, and ΔE is the change in energy consumption.

[0037] Furthermore, the present invention also discloses an enterprise production index monitoring system, which includes: Data acquisition module: used to collect power data, water consumption data, heat / gas energy consumption data and equipment sensor data; Data processing module: realizes the calculation of production trend index, product energy efficiency index, comprehensive energy efficiency index and other indicators; Intelligent diagnosis module: Fault diagnosis and early warning based on Bayesian network; Data storage and verification module: stores data in the blockchain and verifies its credibility, while also performing exception review.

[0038] In practical application, the following steps are specifically included: Step (a), data collection: Collect the company's real-time electricity data, including electricity consumption this month, electricity consumption last month, and electricity consumption in the same period in history; at the same time, collect water consumption, heat / gas energy consumption data and equipment sensor data, including vibration sensors, temperature sensors and pressure sensors.

[0039] Smart meters collect electricity data at intervals of 15 minutes, device sensors collect data every 5 minutes, and water / heat energy consumption data is collected in real time. The raw data is pre-processed by denoising and normalization through edge computing nodes and then transmitted to the cloud platform via the 5G network.

[0040] Step (b), calculate the production trend index: Production trend index = expected month-on-month increase × W3 + current month-on-month increase × W4, in: Expected month-on-month growth = (electricity consumption in the second month of last year / electricity consumption in this month of last year) × 100%; Current month-on-month change = (this month's electricity consumption / last month's electricity consumption) × 100%; W3 and W4 are dynamic weights and satisfy W3+W4=100%.

[0041] In practical applications, the periodic fluctuation intensity is first calculated based on the historical electricity consumption series. The fluctuation intensity is quantified by variance, and the fluctuation intensity is positively correlated with W3.

[0042] If the historical electricity consumption variance is greater than the upper threshold (such as 10,000), W3 is assigned a value of 70%; if it is less than the lower threshold (such as 5,000), W3 is assigned a value of 30%; when it is between the thresholds, W3 is calculated by linear interpolation.

[0043] Step (c), calculate the product energy efficiency index: Product energy efficiency index = (unit energy consumption of standard product / (this month's electricity consumption ÷ output)) × 100%, where the unit energy consumption of standard product is calculated by aggregating the total monthly electricity consumption and total output of multiple manufacturers in the same industry through cloud-based aggregation.

[0044] At 0:00 on the 1st of each month, the cloud server obtains the monthly electricity consumption and output data of at least 10 representative companies in the same industry, generates the industry benchmark value according to the formula standard product unit energy consumption = ∑ (manufacturer's monthly electricity consumption) / ∑ (manufacturer's monthly output), and completes the update.

[0045] Step (d), generate comprehensive energy efficiency index CEI: ,in , Estd, Wstd, and Tstd are pre-stored industry energy consumption benchmarks, and Hscore is the equipment health coefficient.

[0046] The calculation process of the equipment health coefficient Hscore is as follows: First, obtain the vibration deviation value ΔV, temperature deviation value ΔT, and pressure deviation value ΔP and normalize them, then use the formula: Calculation, weights w1, w2, w3 are dynamically allocated based on equipment capacity contribution or maintenance cost.

[0047] Dynamic Weight Adjustment methods include: Load industry preset weight templates, such as the chemical industry ; Generate energy consumption fluctuation correction factor through Sigmoid function; Normalize the corrected weight to meet .

[0048] Step (e), abnormal diagnosis and early warning: When the comprehensive energy efficiency index CEI is lower than a threshold, fault diagnosis is performed.

[0049] First, mutation characteristics of multi-source data are detected, such as power surges >20%, water temperature deviations >5°C, and vibration exceeding the standard >50%. Then, the posterior probability of the fault is calculated based on a Bayesian network using the formula P(fault|evidence) ∝ P(evidence|fault) × P(fault), where the prior probability P(fault) comes from the historical database and the conditional probability table predefines the relationship between evidence and faults. For example, when the power surge is >20%, the probability of the cooling system failure is 0.8; when the water temperature deviation is >5°C, the probability of the cooling system failure is 0.9; and when the vibration exceeds the standard >50%, the probability of the pump failure is 0.85. Finally, the fault type corresponding to the maximum posterior probability is output and a warning signal is triggered.

[0050] Step (f) Data Credibility Verification and Anomaly Review: Real-time electricity data collected by smart meters is stored on the blockchain, and a production digital fingerprint is generated based on the blockchain data. Financial data is compared with the on-chain electricity data through the tax audit interface. If the discrepancy is greater than 3%, a manual review is triggered. Simultaneously, production monitoring indicators are cross-validated with financial output data. If |(Δoutput / Δenergy consumption)-1|>0.3 is met, a manual review process is triggered.

[0051] In order to facilitate those skilled in the art to further understand the present invention, the present invention is further described below with reference to specific implementation cases.

[0052] Case 1: Take the plastic granule production plant of a chemical enterprise as an example.

[0053] Install smart meters (accuracy level 0.5) to collect real-time data on electricity consumption this month (CU), last month (PU), the same period last year (LYC), and next month (LYN), and upload data to the edge computing node every 15 minutes.

[0054] Vibration sensors (accuracy ±0.1 mm / s), temperature sensors (accuracy ±0.5°C), and pressure sensors (accuracy ±1% FS) are deployed on key equipment (such as reactors and pumps) to collect data every 5 minutes.

[0055] Install electromagnetic water meters (accuracy ±0.5%) and heat meters (accuracy ±1.5%) to monitor water consumption (WU) and heat / gas energy consumption (HU) in real time.

[0056] The edge node pre-processes the raw data, including denoising and normalization, and transmits it to the cloud platform via the 5G network; the cloud platform aggregates the monthly electricity consumption and output data of 10 companies in the same industry and calculates the standard product unit energy (SPE).

[0057] Calculate the Production Trend Index (TI): Calculate expected month-on-month growth: EC = (LYN / LYC) × 100% Calculate the current situation year-on-year: CC = (CU / PU) × 100%; Among them, LYN=8000kWh, LYC=7500kWh, EC=106.67%; CU=7800kWh, PU=7200kWh, CC=108.33%.

[0058] Calculate the variance (VAR) of electricity consumption over the past 12 months. If VAR ≥ 10,000 (high volatility), W3 = 70%; if VAR ≤ 5,000 (low volatility), W3 = 30%; intermediate values are calculated using linear interpolation.

[0059] VAR=8000, W3=50%, W4=50%.

[0060] Calculate the production trend index TI=106.67%×50%+108.33%×50%=107.5%.

[0061] Calculate the standard product unit energy consumption (SPE). The total electricity consumption of the same industry is 1.2 million kWh, and the total output is 12,000 tons. SPE=100 kWh / ton.

[0062] Calculate the actual unit energy consumption (UE) of the product, CU=7800kWh, output 75 tons, UE=104kWh / ton.

[0063] Product energy efficiency index EEI = (104 / 100) × 100% = 96.15%.

[0064] Equipment health factor (Hscore), vibration deviation ΔV = (actual value - reference value) / reference value, temperature deviation ΔT, pressure deviation ΔP are similar.

[0065] ΔV=0.1, ΔT=0.05, ΔP=0.08, weights w1=0.5 (capacity contribution), w2=0.3, w3=0.2.

[0066] ; Dynamic Weight ( ): Load the chemical industry template ( ).

[0067] According to the energy consumption fluctuation coefficient (such as power fluctuation coefficient = 1.2), the correction factor is generated through the Sigmoid function:

[0068] Power fluctuation coefficient 1.2, f=0.731, ; After normalization .

[0069] Comprehensive energy efficiency index: ; Assuming water / heat energy consumption meets the standards: CEI=96.15%×25%+100%×30%+100%×30%+14.7%=98.79%.

[0070] Adjustable weights (W1 / W2): Enterprises set W1=60% (focus on trends) and W2=40% (focus on energy efficiency) based on management needs.

[0071] Production monitoring indicator: PMI = 107.5% × 60% + 96.15% × 40% = 103.36%. If the preset threshold T = 100%, the PMI meets the standard and there is no warning.

[0072] In actual fault diagnosis, when the CEI is lower than the threshold, if CEI < 90%, detection is triggered: Power surge > 20%, such as a sudden increase in CU from 7200kWh to 8640kWh; Water temperature deviation> 5℃, such as cooling water temperature rises from 30℃ to 36℃; Vibration exceeds the standard by >50%, such as the vibration value increases from 2mm / s to 3.2mm / s.

[0073] Using Bayesian network reasoning, the conditional probability table (CPT) is shown in Table 1 below: Table 1 is the conditional probability table:

[0074] Calculation of the posterior probability: P(cooling system failure | E) ∝ 0.8 × 0.9 × P(cooling system failure). Assuming the prior probability P(cooling system failure) = 0.1 and the probability of other failures P = 0.9, the posterior probability = (0.72 × 0.1) / (0.72 × 0.1 + 0.05 × 0.9) = 0.616.

[0075] Diagnostic result: The cooling system failure probability is 61.6%, triggering a work order for maintenance.

[0076] Blockchain evidence is used for data credibility verification and exception review. Smart meter data is encrypted using SHA-256 hashing and written to a blockchain, such as Hyperledger Fabric, to generate an unalterable digital fingerprint. This data is compared monthly with tax declaration data. If the discrepancy exceeds 3%, for example, if the on-chain electricity consumption is 7800 kWh and the tax declaration is 7500 kWh, manual verification is triggered.

[0077] When performing cross validation, the energy efficiency output ratio (R): , Output increased from 70 tons to 75 tons, ΔQ=7.14%, electricity consumption increased from 7200kWh to 7800kWh, ΔE=8.33%, R=0.857.

[0078] Abnormal determination: , no manual review was triggered.

[0079] Through dynamic weight adjustment and CEI monitoring, the energy consumption per unit product has been reduced from 104kWh / ton to 100kWh / ton, saving approximately RMB 120,000 in electricity costs annually. The Bayesian network diagnosis accuracy rate reaches 90%, detecting cooling system hidden dangers 72 hours in advance and avoiding production losses. Blockchain evidence storage reduces the risk of data tampering, improves tax comparison efficiency, and uses real-time monitoring of the equipment health coefficient Hscore, reducing maintenance costs by 30%, and increasing the spare parts inventory turnover rate to 5.8 times / year.

[0080] Case 2: For hardware deployment, a chemical company installed smart meters for electricity data collection. They also deployed vibration, temperature, and pressure sensors on key equipment like reactors and pumps. They also installed electromagnetic water meters and heat meters to monitor water and heat / gas consumption. The collected data was pre-processed by edge computing nodes and then transmitted to the cloud platform.

[0081] In the indicator calculation stage, taking a certain month as an example, the electricity consumption (LYN) in the next month of last year was 8000kWh, and the electricity consumption (LYC) in this month of last year was 7500kWh. The expected month-on-month growth is: (EC)=(8000 / 7500)×100%=106.67%; The electricity consumption (CU) this month is 7800kWh, and the electricity consumption (PU) last month was 7200kWh. The current situation is: (CC)=(7800 / 7200)×100%=108.33%.

[0082] The variance of electricity consumption in the past 12 months is calculated to be 8000. The dynamic weight calculation rule is used to obtain W3=50%, W4=50%, and then the production trend index is: (TI)=106.67%×50%+108.33%×50%=107.5%.

[0083] Standard product unit energy consumption (SPE) is calculated by obtaining data from 10 companies in the same industry through the cloud. Assuming the total electricity consumption of the industry is 1.2 million kWh and the total output is 12,000 tons, then SPE = 100 kWh / ton. The company's output this month is 75 tons. The actual product unit energy consumption is: (UE) = 7800 / 75 = 104kWh / ton, product energy efficiency index: (EEI)=(100 / 104)×100%=96.15%.

[0084] When calculating the equipment health factor, obtain the vibration, temperature, and pressure deviation values and normalize them. Assuming ΔV=0.1, ΔT=0.05, ΔP=0.08, weights w1=0.5, w2=0.3, w3=0.2, the calculated Hscore=0.98. Load the preset weight template for the chemical industry, and adjust the weights using the Sigmoid function based on the energy consumption fluctuation coefficient to obtain The comprehensive energy efficiency index CEI is calculated to be 98.79%.

[0085] If the CEI in a particular month falls below a preset threshold of 90%, the system detects sudden changes such as power surges and water temperature deviations. Using a Bayesian network to calculate posterior probabilities, the system diagnoses the cooling system as having the highest probability of failure, triggering an alert and generating a repair work order. Furthermore, blockchain-based evidence storage and data cross-validation ensure data accuracy and reliability, triggering manual review if anomalies are detected.

[0086] Through the implementation of the present invention, the chemical enterprise has achieved accurate monitoring and optimization of production indicators, reduced energy consumption per unit product, reduced losses due to equipment failure, and achieved significant economic and management benefits.

[0087] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0088] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. It should be pointed out that any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. The enterprise production index monitoring method based on power data is characterized by: Specifically, it includes the following steps: Step 1: Collect enterprise power data in real time; Step 2: Dynamically allocate weights based on the historical electricity consumption fluctuation intensity, and calculate the production trend index TI; Step 3: Calculate the product energy efficiency index EEI according to the industry benchmark; Step 4: Integrate TI, EEI and the equipment health coefficient Hscore to generate the comprehensive energy efficiency index CEI, where Hscore is dynamically calculated based on the equipment sensor deviation value; Step 5: When CEI is lower than the threshold, perform fault diagnosis through the Bayesian network and trigger an early warning.

2. The enterprise production index monitoring method based on power data according to claim 1, wherein: In Step 1, collect the real-time power data of the enterprise, including the current month's electricity consumption CU, the previous month's electricity consumption PU, the electricity consumption LYC in the same month last year, and the electricity consumption LYN in the next month last year; In Step 2, specifically include the following steps to calculate the production trend index TI; Calculate the expected month-on-month ratio: EC = (LYN / LYC) × 100% Calculate the current month-on-month ratio: CC = (CU / PU) × 100% Dynamically allocate weights W3 and W4, and satisfy W3 + W4 = 100%; Generate TI = EC × W3 + CC × W4; In Step 3, calculate the product energy efficiency index EEI specifically as follows; Obtain the industry standard product unit energy consumption: SPE = total monthly electricity consumption of all manufacturers in the same industry ÷ total output; Calculate the actual product unit energy consumption: UE = CU ÷ the output Q in the current month; Generate EEI = (SPE / UE) × 100%; In Step 4, generate the production monitoring index: PMI = TI × W1 + EEI × W2, where W1 + W2 = 100%.

3. The method for monitoring enterprise production indicators based on power data according to claim 2, characterized in that: The assignment method of the dynamic weight W3 in Step 2 is as follows. Based on the electricity consumption data in the past 12 months, calculate its variance VAR as the fluctuation intensity index VI; VI = VAR(E1, E2, …, E12); Among them, Ei is the electricity consumption in the i-th month, VI is the fluctuation intensity, and VAR is the variance calculation function; Determine the dynamic weight W3 according to the fluctuation intensity (VI): If VI ≥ high threshold V1, then W3 = 70%, paying more attention to the expected month-on-month ratio; If VI ≤ low threshold V2, then W3 = 30%, paying more attention to the current month-on-month ratio; If low threshold V2 < VI < high threshold V1, then calculate through linear interpolation: ； According to W3 + W4 = 100%, calculate W4 = 100% - W3.

4. The enterprise production index monitoring method based on power data according to claim 2 is characterized in that: The calculation method of the standard product unit energy consumption is specifically as follows: S101: Data collection. Obtain the monthly data of at least 10 representative enterprises in the same industry through the cloud server, including: The monthly total output (Q1, Q2, …, Qn) of each enterprise; The total monthly electricity consumption (E1, E2, …, En); S102: Calculate the industry benchmark: Calculate the total industry electricity consumption: ∑E = E1 + E2 + … + En; Calculate the total industry output: ∑Q = Q1 + Q2 + … + Qn; Calculate standard product specific energy (SPE): ; S103: Data update. Automatically update the SPE value every month to ensure the timeliness of the benchmark.

5. The enterprise production index monitoring method based on power data according to claim 2, wherein: Step 1 also includes collecting water consumption, heat / gas energy consumption data and equipment sensor data, wherein the equipment sensors include vibration sensors, temperature sensors and pressure sensors; On the basis of the original power data, the following data are collected: Water consumption data: total water consumption (WU) and water consumption per unit product (WUP) for the month; Heat / gas energy consumption data: total heat / gas volume (HU) for the month, heat / gas consumption per unit product (HUP); Device sensor data: Vibration sensor: real-time vibration value (V), reference vibration value (Vstd); Temperature sensor: real-time temperature value (T), reference temperature value (Tstd); Pressure sensor: real-time pressure value (P), reference pressure value (Pstd); Step 4 is replaced by generating the comprehensive energy efficiency index CEI: CEI=(Estd / Eactual×α)+(Wstd / Wactual×β)+(Tstd / Tactual×γ)+(Hscore×δ); Among them, α, β, γ, and δ are dynamic weights, α+β+γ+δ=1; Estd, Wstd, and Tstd are pre-stored industry energy consumption benchmarks, and Eactual, Wactual, and Tactual are actual energy consumption values.

6. The enterprise production index monitoring method based on power data according to claim 5 is characterized in that: The calculation process of the equipment health coefficient Hscore is as follows: ； ΔV=(Vactual-Vstd) / Vstd; ΔT=(Tactual-Tstd) / Tstd; ΔP=(Pactual-Pstd) / Pstd; Among them: Vactual, Tactual, Pactual are actual measured values, Vstd, Tstd, Pstd are standard values.

7. The method for monitoring enterprise production indicators based on power data according to claim 6, characterized in that: The adjustment methods of dynamic weights α, β, γ, and δ include: Calculate the energy consumption fluctuation coefficients Evar, Wvar, Tvar, and Hvar, By correcting the function Generate scaling factors; according to Calculate intermediate weights; The final weights α, β, γ and ; The details are as follows: ； in: The preset weights are for the industries, and Evar, Wvar, Tvar, and Hvar are energy consumption fluctuation coefficients.

8. The method for monitoring enterprise production indicators based on power data according to claim 5, characterized in that: When the CEI is lower than the threshold, perform fault diagnosis. The specific steps are as follows: P(F|E)∝P(E|F)×P(F); Fdetected=argmax[P(F|E)]; Where: F is the fault type, E is the mutation feature of the evidence monitoring data, P(F|E) is the posterior probability of fault type F under evidence E, argmax[P(F|E)] is the fault type with the largest posterior probability as the diagnosis result Fdetected, P(E|F) is the conditional probability of evidence E when fault F occurs, and P(F) is the prior probability of the fault.

9. The method for monitoring enterprise production indicators based on power data according to claim 8, characterized in that: Conditional probabilities include: P(CF|EP>20%)=0.8; P(CF|TD>5℃)=0.9; P(PF|VO>50%)=0.85; Among them: CF is cooling system failure, EP is power surge, TD is water temperature deviation, PF is pump failure, and VO is vibration exceeding the standard.

10. The method for monitoring enterprise production indicators based on power data according to any one of claims 1 to 9, characterized in that: It also includes data credibility verification steps: ； Vresult=Compare(Fdata,BCdata); Among them: DFP is the production digital fingerprint, refers to the real-time electricity data collected by the enterprise for each month, Fdata is the financial data, BCdata is the blockchain data, and Vresult is the difference result of comparing the financial data Fdata and the blockchain data BCdata; the Compare function calculates the relative difference between the key fields of the financial data Fdata and the blockchain data BCdata.

Images