Slurry treatment operation control calculation method for slurry balance shield

Abstract

The invention provides a slurry treatment operation control calculation method for a slurry balance shield. The method comprises the following steps: S1, data acquisition; a weighing sensor, a flow meter, a densimeter, an electricity meter and a PLC monitoring device are configured in a shield construction system. The method comprises the following steps: S2, data preprocessing; s3, cost calculation; s4, summarizing the total cost; the system performs cost classification collection and numerical value accumulation according to a preset time period according to the completed various single cost calculation results to form a unified total cost data table, and performs structured summarization on various cost factors; s5, cost analysis and feedback; and continuous monitoring and feedback adjustment of the cost structure are realized. The system combines a dynamic correction mechanism and a risk assessment algorithm to carry out trend prediction and anomaly identification on cost fluctuation, so that the problem of high cost or low efficiency can be found in time, managers are guided to formulate targeted optimization measures, and thus the cost management precision and response efficiency of slurry treatment operation are improved.

Description

Technical Field

[0001] This invention relates to the field of shield tunneling control technology, specifically to a calculation method for the operation and control of slurry treatment in slurry balance shield tunnels. Background Technology

[0002] In slurry-balanced shield tunneling, the slurry treatment system is a core component ensuring the stability of the shield excavation face and efficient separation of excavated soil. Its operational status directly impacts construction safety, progress, and cost. However, existing slurry treatment systems face multiple technical challenges, making it difficult to balance operational control and cost accounting. 1. Low control accuracy due to multi-parameter coupling interference: Slurry treatment involves multiple physical property parameters such as density, viscosity, and sand content. These parameters are strongly coupled (e.g., increased sand content leads to increased viscosity, which in turn increases pump energy consumption). Existing systems often use manual adjustment of a single parameter (e.g., adding bentonite only based on density), without considering the linkage effect between parameters, which easily leads to "overdosing" or "adjustment lag"—for example, in one project, due to the lack of correlation with viscosity data, excessive addition of bentonite increased the cost of slurry treatment by 12% per ton, while also causing the risk of pipeline blockage.

[0003] 2. Poor sensor data reliability: The mud processing environment involves high vibration, high humidity, and corrosive media (such as chemical additives in the mud), which makes sensors prone to drift or failure. Existing systems lack regular calibration mechanisms and data verification logic. For example, in one project, the density meter had an error of ±0.03 g / cm³ due to lack of calibration, resulting in a 15% deviation in bentonite usage calculations, directly leading to distorted cost accounting. Simultaneously, the sampling times of multiple devices are not synchronized (with deviations reaching 5-10 minutes), making it impossible to establish a time-series correspondence between "equipment operation - material consumption - cost".

[0004] 3. Cost accounting is disconnected from operating conditions: Existing cost calculations are mostly based on static data (such as fixed raw material unit prices and uniform depreciation rates), without considering the impact of dynamic changes in operating conditions. For example, the sand content of mud in gravel strata is 8-10% higher than that in clay strata, leading to a 30% increase in filter aid usage and a 25% increase in filter press energy consumption. However, existing models do not introduce operating condition correction factors and still calculate based on fixed parameters, resulting in cost prediction deviations exceeding 20%. In addition, equipment depreciation is not linked to actual operating load; idle equipment and full-load equipment are depreciated according to the same standard, further exacerbating accounting errors.

[0005] 4. Lack of closed-loop feedback control mechanism: The existing system can only perform post-event statistics on cost data and cannot adjust operating parameters in reverse according to cost fluctuations. For example, when electricity costs rise abnormally, it is impossible to quickly determine whether it is "pump failure causing increased energy consumption" or "abnormal mud viscosity causing increased equipment load". Manual investigation is required, with a response time of up to 4-6 hours, missing the optimal adjustment opportunity and resulting in additional cost losses.

[0006] The aforementioned technical challenges are not simply economic issues, but rather stem from the lack of coordinated control over "mud physical properties, equipment operating status, and cost factors," resulting in low system efficiency and poor cost controllability. Therefore, there is an urgent need for a technical solution that integrates multi-parameter coordinated control, data reliability assurance, and dynamic cost accounting to address the technical deficiencies of the existing system. Summary of the Invention

[0007] The technical problem to be solved by this invention is to provide a calculation method for the operation control of mud treatment in slurry balance shield tunneling. This method solves the technical problems of low control accuracy, poor reliability of sensor data, disconnect between cost accounting and working conditions, and lack of closed-loop feedback control mechanism caused by multi-parameter coupling interference in the current construction process. It enables precise parameter control and dynamic cost optimization of the mud treatment system.

[0008] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: A calculation method for the operation and control of slurry treatment in slurry-balanced shield tunneling machines, the method comprising: S1. Data Acquisition: Weighing sensors, flow meters, density meters, electricity meters, and PLC monitoring devices are installed on material conveying devices, liquid storage tanks, pumps, and chemical dosing devices at key process nodes in the shield tunneling system, such as the pre-screening unit, slurry separation unit, slurry preparation unit, and waste slurry treatment unit. These devices are used to collect real-time information on the addition amount and flow rate of materials such as bentonite, slurry preparation agent, thickener, filter aid, and slag, as well as the power consumption data of various pumps, motors, and hydrocyclones. Through system integration, ledger management, and manual input, data related to equipment depreciation, including the original value of the equipment, service life, current usage time, or remaining usable years, are collected. At the same time, a model for operating cost accounting is obtained. S1 also includes sensor calibration and synchronization control: every set number of days, all sensors are calibrated using the national metrological standard material comparison method, wherein the calibration error of the weighing sensor must be ≤ ±0.1kg, the calibration error of the flow meter must be ≤ ±0.5%FS under full scale conditions, the calibration error of the density meter must be ≤ ±0.005g / cm³, and the calibration error of the electricity meter must be ≤ ±0.2 grade; the time synchronization of all acquisition devices is achieved through the GPS time synchronization module, with a time synchronization error of ≤1ms, ensuring the consistency of multi-source data in the time dimension; S2, Data Preprocessing; S3. Cost Calculation; S4. Total Cost Summary: Based on the completed individual cost calculation results, the system classifies, collects, and accumulates material costs, electricity costs, spare parts costs, labor costs, equipment depreciation, etc., according to a preset time period to form a unified total cost data table, and performs structured summary of various cost factors. S5. Cost analysis and feedback: To achieve continuous monitoring and feedback adjustments of the cost structure.

[0009] The specific process of S2 mentioned above is as follows: The data collected in S1 is multi-source heterogeneous data. For multi-source heterogeneous data, wavelet transform is used to denoise the text, and breadth-first search is combined to extract key indicators to achieve the required structure and standardization. At the same time, Z-score standardization, isolated forest outlier identification and dynamic time warping are performed on numerical data to unify the data format, remove abnormal interference, and align different time series data to ensure data consistency and usability.

[0010] The specific process of S5 mentioned above is as follows: After the system completes the total cost summary, it generates a cost analysis report at a set period, performs statistical analysis and trend judgment on various cost data, and identifies cost items with abnormal fluctuations or exceeding preset thresholds; at the same time, it combines the operation data of each process to locate potential high-consumption points or efficiency deviations, so as to realize continuous monitoring and feedback adjustment of the cost structure. S5 also includes a mud properties-cost linkage feedback: based on the correlation model between cost analysis results and mud density and viscosity data, when the mud density is lower than ρ... 标准 Furthermore, if the cost of bentonite does not exceed the limit, a control command to "increase the amount of bentonite added" is automatically generated. The command includes a specific adjustment amount, based on the "density-bentonite dosage" mapping formula fitted from historical data: ΔQ=k×(ρ 标准 -ρ 实际 ), where k is an adjustment coefficient, ranging from 1.2 to 1.5 kg / (g·cm³). -3) , ρ 标准 =1.1g / cm³; When the slurry viscosity is higher than 30s and the cost of the thickener exceeds the limit, a compound instruction of "reducing the amount of thickener added + adjusting the speed of the mixing pump" is automatically generated. The adjustment range of the mixing pump speed is positively correlated with the viscosity deviation value, and the working condition coefficient β in the cost accounting model is updated simultaneously.

[0011] The process of obtaining the operating cost accounting model from S1 mentioned above also includes: S1-1. Definition of Data Items: The data types required for cost accounting are categorized and defined, and their collection methods and units are specified, serving as the basis for subsequent data collection and analysis. Data items mainly include four categories: First, equipment power consumption data, primarily collecting the electricity consumption of various operating equipment, collected via electricity meters, with units of kW·h; second, material consumption data, involving the usage of raw materials such as bentonite, soda ash, thickeners, and filter aids, collected through weighing equipment, PLC systems, or manual recording, with units including kg or m³; third, auxiliary analysis indicators, including mud characteristic data and equipment operating status data, used to support cost analysis and operational efficiency assessment; and fourth, manually entered cost data, mainly including economic parameters such as raw material unit prices, labor rates, and equipment depreciation rates, which are regularly maintained and entered by management personnel. Auxiliary analysis indicators also include equipment operating temperature data: by installing temperature sensors on the casing of pumps and motors, real-time equipment temperature is collected. When the temperature exceeds 80℃, the "temperature-power consumption loss coefficient" γ is automatically calculated by associating it with power consumption data, where γ = 1 + 0.01 × (T 实际 -80), T 实际 The real-time temperature of the equipment is calculated, and γ is incorporated into the cost accounting model to adjust the equipment's electricity costs. S1-2, Data Standard Definition: Standardize the accuracy requirements, time synchronization methods, and data structure of the collected data; sensor data must meet the set measurement accuracy standards to avoid interference from errors affecting subsequent calculation results; all data should be aligned by timestamps, and the sampling frequency should meet the minimum acquisition requirements to ensure consistent synchronization of multi-source data in the time dimension; collected data should be stored in a unified structured format, including field names, units, and verification information, providing standardized support for subsequent data parsing, transmission, and processing.

[0012] The cost calculation in S3 above specifically includes: S3-1, Definition of Cost Items: Classify and standardize various cost elements involved in cost accounting, and clarify the accounting scope and attribution standards; Cost items are mainly divided into two categories: one is direct costs, including raw material consumption, energy expenditure and on-site labor costs, which are directly related to the actual input of resources during the construction process; the other is indirect costs, including equipment depreciation, daily maintenance costs and potential environmental penalties, which are expenditures that cannot be directly attributed to a certain operation link. S3-2, Baseline Cost Calculation: Based on the collected information on raw material usage, unit price, labor costs, and equipment depreciation, a standardized baseline cost calculation model is established. By summing the products of various raw material usages and their unit prices, adding the product of labor rates and working hours, and factoring in equipment depreciation costs, the basic cost value for the construction period is calculated. The calculation formula is as follows: ; Among them, Q i Let P be the amount of the i-th raw material used. i Let L be the unit price of the raw material, L be the labor cost rate, T be the labor hours, and D be the labor cost. equip Equipment depreciation costs; Equipment depreciation cost D equip The calculation introduces the "equipment operating load correction factor" δ: δ = actual operating time / rated operating time, and the corrected D equip = (Original value of equipment / Estimated useful life / 365) × δ, to ensure that the equipment depreciation cost matches the actual usage intensity and avoid accounting deviations caused by excessive depreciation of idle equipment; S3-3, Dynamic Correction Factor: A dynamic adjustment mechanism is introduced based on the baseline cost calculation to improve the timeliness and adaptability of cost accounting. By introducing the historical deviation rate α and the real-time operating condition coefficient β, the baseline cost is dynamically corrected, and the predicted cost C is calculated. pred The corrected formula is as follows: ; Among them, α is derived from the regression analysis results of historical data and reflects the trend of statistical deviation; β is the working condition correction factor that reflects the current changes in mud characteristics and is used to dynamically adjust the cost impact caused by changes in physical states such as sand content and viscosity. S3-4, Risk Assessment: The predicted cost is calculated multiple times using Monte Carlo simulation to assess the probability of exceeding the budget, thereby quantifying cost risks during construction. The system is based on dynamically corrected predicted costs. The sampling is repeated within a set number of simulations N, and it is determined whether the result exceeds the set budget value B each time; finally, the probability of cost overrun P is calculated using the following formula. over : ; Where I(⋅) is an indicator function used to determine whether the current simulation result exceeds the budget.

[0013] The sensors configured in S1 above have a one-to-one correspondence with the collected data: the weighing sensor is used to collect data on the amount of bentonite, slurry agent, thickener, and filter aid added, in kg; the flow meter is used to collect data on the flow rate of slag, clean water, and fresh slurry, in m³ / h; the densitometer is used to collect data on the density of the mud, in g / cm³; the electricity meter is used to collect data on the power consumption of various pumps, motors, and hydrocyclones, in kW·h; and the PLC monitoring device is used to collect data on the amount of flocculant added to the filter press and centrifuge, in ml.

[0014] The specific process of outlier identification in isolated forests is as follows: anomaly score is calculated for each numerical data sample based on the isolated forest algorithm. An anomaly score threshold of 0.6 is set. When the anomaly score of a sample is greater than 0.6, the sample is determined to be an outlier and is directly removed. The specific process of dynamic time warping is as follows: the sampling time of the power consumption data of the pre-screening unit is used as the reference time axis. The time series data of the mud-water separation unit, the slurry preparation unit, the pulping unit, and the waste slurry treatment unit are adjusted by the dynamic time warping algorithm so that the timestamp deviation of all time series data does not exceed 1 minute.

[0015] The specific equipment and data collection requirements for the above-mentioned power consumption data are as follows: power of the pre-screening unit, pump motor power of the first-stage hydrocyclone unit, vibration power of the dewatering screen unit, pump motor power of the second-stage hydrocyclone unit, power of the clear water pump motor in the clear water tank, power of the slurry mixing pump motor and stirring pump motor in the slurry mixing tank, power of the return slurry pump motor in the return slurry tank, power of the waste slurry pump motor in the waste slurry tank, power of the filter press pump motor in the waiting-to-press tank, power of the stirring pump motor, filtrate water pump motor and main motor power in the filter press, and power of the centrifuge, centrifuge power, two flocculant pump motors, and clear water pump motor power in the centrifuge; the sampling frequency for all equipment power consumption data shall not be less than once per minute.

[0016] The specific materials and collection methods corresponding to the above-mentioned material consumption data are as follows: the sludge removal volume of the sedimentation tank is collected manually, in m³; the pulping agent of the pulping system is collected by weighing equipment, in kg; the bentonite in the new pulp tank is collected by weighing equipment, and the feed rate of the new pulp pump and the addition rate of the shear pump are collected by flow meters, in kg and m³ / h respectively; the thickener in the slurry conditioning tank and the return slurry tank is collected by weighing equipment, in kg; the filter aid in the waste slurry tank and the waiting press tank is collected by weighing equipment, in kg; the flocculant addition rate of the filter press is recorded by the PLC system, in ml; and the flocculant addition rate and the clean water addition rate of the centrifuge are recorded by the PLC system, in ml.

[0017] The above historical deviation rate The calculation process is as follows: Select historical data from the last three construction cycles, calculate the difference between "historical actual cost - historical benchmark cost" in each cycle, and then take the arithmetic mean of the ratio of the difference to the historical benchmark cost of the corresponding cycle to obtain the historical deviation rate. If the historical actual cost is greater than the historical benchmark cost, Take a positive value; if the historical actual cost is less than the historical benchmark cost, Take the negative value; The calculation precision is retained to 4 decimal places.

[0018] The above-mentioned working condition coefficients The determination is based on the standard sand content of the mud. Standard viscosity Based on the actual sand content Each higher At 1%, Increase by 0.02 from 1.0; when the actual sand content... Every lower At 1%, Reduce by 0.01 from 1.0; when the actual viscosity Each higher At 2s, Increase by 0.03 from 1.0; when the actual viscosity Every lower At 2s, Reduced by 0.015 from 1.0; The calculation result is rounded to three decimal places.

[0019] The above Monte Carlo simulation sets the number of simulations. No less than 1000 times; for each sampling, the amount of raw materials used... Labor cost rate Equipment depreciation costs Apply random fluctuations of ±5%, and ensure the sampling process follows a normal distribution; after the simulation, if the probability of cost overrun is... If the percentage exceeds 20%, a cost warning will be triggered.

[0020] The preset time periods in S4 include daily, weekly, and monthly periods, and user-defined periods are also supported. When classifying and collecting costs, material costs are subdivided into "bentonite cost, slurry agent cost, thickener cost, filter aid cost, and flocculant cost", electricity costs are subdivided into "electricity costs of equipment in each process", and spare parts costs are subdivided into "cost of vulnerable parts and cost of core components", ensuring that the collection error of each type of cost does not exceed 2%.

[0021] The aforementioned preset thresholds are set as follows: based on historical cost data from the past 6 months, the historical average value of each cost item is calculated. with standard deviation The preset threshold range is When the actual value of a certain cost item exceeds the specified range, the system automatically marks it as an "abnormal cost item" and notes the corresponding process, equipment, and data acquisition time in the cost analysis report.

[0022] The application process of the aforementioned auxiliary analysis indicators is as follows: Correlation analysis is performed between mud density and viscosity data and raw material usage data. When the mud density is below 1.1 g / cm³, a suggestion to "increase the amount of bentonite added" is automatically generated; when the mud viscosity is above 30s, a suggestion to "reduce the amount of thickener added" is automatically generated. The suggestions include specific adjustment amounts: a "characteristic-usage" mapping formula fitted based on historical data. ,in To adjust the coefficient, These are actual characteristic values. The standard characteristic values ​​are used, and the cost prediction model is updated synchronously.

[0023] The slurry treatment operation control calculation method for slurry balance shield tunneling mentioned in this invention constructs a multi-dimensional cost model covering raw materials, electricity, labor, equipment depreciation, and other elements through real-time acquisition and dynamic analysis of key cost factors of the slurry treatment system. This model clearly identifies the cost distribution and changing trends at each stage, providing a precise data foundation for cost control during construction. The system combines a dynamic correction mechanism and risk assessment algorithm to predict cost fluctuations and identify anomalies, helping to promptly detect cost overruns or inefficiencies and guiding managers to develop targeted optimization measures, thereby improving the accuracy and response efficiency of slurry treatment operation cost management. The operation control in this application has the following beneficial effects: 1. Sensor calibration and synchronization control: By clarifying the calibration cycle, error standard, and GPS time synchronization, the technical problems of sensor drift and data asynchrony in the slurry treatment environment are solved. This belongs to the technical characteristics of the "measurement control" field, conforms to objective physical laws, and sensor accuracy decays with usage time, requiring periodic calibration; data from multiple devices need to be synchronized in time to establish causal relationships.

[0024] 2. Data integrity verification and completion mechanism: Through three-dimensional verification and moving average completion, the problem of accounting interruption caused by missing data is solved. This is a technical feature in the field of "data processing". The error control of the completion algorithm is ≤2%, which meets the objective accuracy requirements of data processing.

[0025] 3. Mud Characteristics-Cost Linkage Feedback: Constructing a closed loop of "cost-parameter-control command" through a quantitative mapping formula (ΔQ=k×(ρ)). 标准 -ρ 实际 This achieves multi-parameter coordinated adjustment, solves the problem of parameter coupling interference, and conforms to the objective law that "physical properties determine the amount of material used." The slurry density needs to be adjusted by the amount of bentonite used, and the viscosity needs to be controlled by the coordinated control of the thickener and the stirring speed.

[0026] 4. Equipment temperature-power consumption loss coefficient: Data is collected by temperature sensors to quantify the additional energy consumption caused by overheating. This conforms to the electrical law that "increased equipment temperature increases resistance, leading to increased energy consumption," and is a technical feature of "equipment operation status monitoring."

[0027] 5. Equipment operating load correction factor: This factor links depreciation costs to actual operating time, conforms to the mechanical principle that "equipment wear and tear is positively correlated with usage intensity," and solves the technical defects of traditional static depreciation. It is a technical feature in the field of "equipment management." Attached Figure Description

[0028] The present invention will be further described below with reference to the accompanying drawings and embodiments: Figure 1 This is a schematic diagram of the overall process of the calculation method for the operation control of mud treatment in a slurry-balanced shield tunneling machine according to the present invention. Detailed Implementation

[0029] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and embodiments.

[0030] Example 1: A calculation method for the operation and control of slurry treatment in slurry-balanced shield tunneling machines, the method comprising: S1. Data Acquisition: Weighing sensors, flow meters, density meters, electricity meters, and PLC monitoring devices are installed on material conveying devices, liquid storage tanks, pumps, and chemical dosing devices at key process nodes in the shield tunneling system, such as the pre-screening unit, slurry separation unit, slurry preparation unit, and waste slurry treatment unit. These devices are used to collect real-time information on the addition amount and flow rate of materials such as bentonite, slurry preparation agent, thickener, filter aid, and slag, as well as the power consumption data of various pumps, motors, and hydrocyclones. Through system integration, ledger management, and manual input, data related to equipment depreciation, including the original value of the equipment, service life, current usage time, or remaining usable years, are collected. At the same time, a model for operating cost accounting is obtained. S1 also includes sensor calibration and synchronization control: all sensors are calibrated every 15 days using the national metrological standard material comparison method. The calibration error of the weighing sensor must be ≤ ±0.1kg, the calibration error of the flow meter must be ≤ ±0.5%FS under full scale conditions, the calibration error of the density meter must be ≤ ±0.005g / cm³, and the calibration error of the electricity meter must be ≤ ±0.2 grade. Time synchronization of all data acquisition devices is achieved through a GPS time synchronization module, with a time synchronization error of ≤1ms, ensuring the consistency of multi-source data in the time dimension. This step can solve the data reliability problem caused by sensor drift and provide an accurate data foundation for subsequent cost accounting. S2, Data Preprocessing; S3. Cost Calculation: Based on high-quality data collected and preprocessed, a standard cost model is constructed according to raw material usage, power consumption, labor costs, and equipment depreciation. Combined with the energy consumption change curves of each link, the dynamic loss of sub-items is further quantified by calculating the energy consumption difference and energy consumption ratio between adjacent levels of different processes, forming a cost loss curve oriented towards time series. Historical deviation correction factors and operating condition coefficients are introduced to dynamically adjust costs, and the cost accounting results for each time period and each functional module are comprehensively output. S4. Total Cost Summary: Based on the completed individual cost calculation results, the system classifies and collects material costs, electricity costs, spare parts costs, labor costs, equipment depreciation, etc., according to preset time periods (such as daily, weekly, monthly) and accumulates the values ​​to form a unified total cost data table, and performs structured summary of various cost factors. S5. Cost analysis and feedback: To achieve continuous monitoring and feedback adjustments of the cost structure.

[0031] The specific process of S2 mentioned above is as follows: The data collected in S1 is multi-source heterogeneous data. For multi-source heterogeneous data, wavelet transform is used to denoise the text, and breadth-first search is combined to extract key indicators to achieve the required structure and standardization. At the same time, Z-score standardization, isolated forest outlier identification and dynamic time warping are performed on numerical data to unify the data format, remove abnormal interference, and align different time series data to ensure data consistency and usability.

[0032] The specific process of S5 mentioned above is as follows: After the system completes the total cost summary, it generates a cost analysis report at a set period, performs statistical analysis and trend judgment on various cost data, and identifies cost items with abnormal fluctuations or exceeding preset thresholds; at the same time, it combines the operation data of each process to locate potential high-consumption points or efficiency deviations, so as to realize continuous monitoring and feedback adjustment of the cost structure. S5 also includes a mud properties-cost linkage feedback: based on the correlation model between cost analysis results and mud density and viscosity data, when the mud density is below 1.1 g / cm³ and the bentonite cost does not exceed the limit, a control command to "increase the amount of bentonite added" is automatically generated. The command includes a specific adjustment amount, based on the "density-bentonite dosage" mapping formula fitted from historical data: ΔQ=k×(ρ 标准 -ρ 实际 ), where k is an adjustment coefficient, ranging from 1.2 to 1.5 kg / (g·cm³). -3 ), ρ 标准=1.1g / cm³; When the slurry viscosity exceeds 30s and the cost of the thickener exceeds the limit, a combined instruction of "reducing the amount of thickener added + adjusting the speed of the mixing pump" is automatically generated. The adjustment range of the mixing pump speed is positively correlated with the viscosity deviation value. For every 2s increase in viscosity, the speed is reduced by 5r / min, and the working condition coefficient β in the cost accounting model is updated simultaneously. This step constructs a closed-loop control of "cost analysis - parameter adjustment - model optimization" to solve the problems of multi-parameter coupling interference and feedback lag.

[0033] The process of obtaining the operating cost accounting model from S1 mentioned above also includes: S1-1. Definition of Data Items: The data types required for cost accounting are categorized and defined, and their collection methods and units are specified, serving as the basis for subsequent data collection and analysis. Data items mainly include four categories: First, equipment power consumption data, primarily collecting the electricity consumption of various operating equipment, collected via electricity meters, with units of kW·h; second, material consumption data, involving the usage of raw materials such as bentonite, soda ash, thickeners, and filter aids, collected through weighing equipment, PLC systems, or manual recording, with units including kg or m³; third, auxiliary analysis indicators, including mud characteristic data and equipment operating status data, used to support cost analysis and operational efficiency assessment; and fourth, manually entered cost data, mainly including economic parameters such as raw material unit prices, labor rates, and equipment depreciation rates, which are regularly maintained and entered by management personnel. S1-1 also includes auxiliary analysis indicators such as equipment operating temperature data: by installing temperature sensors on the casing of pumps and motors, real-time equipment temperature is collected. When the temperature exceeds 80℃, the "temperature-power consumption loss coefficient" γ is automatically calculated by associating it with power consumption data, where γ = 1 + 0.01 × (T 实际 -80), T 实际 The real-time temperature of the equipment is calculated, and γ is incorporated into the cost accounting model to correct the equipment's electricity costs. This step can quantify the additional energy consumption caused by equipment overheating and solve the problem of cost accounting being disconnected from the actual operating status of the equipment. S1-2, Data Standard Definition: Standardize the accuracy requirements, time synchronization methods, and data structure of the collected data; sensor data must meet the set measurement accuracy standards to avoid interference from errors affecting subsequent calculation results; all data should be aligned by timestamps, and the sampling frequency should meet the minimum acquisition requirements to ensure consistent synchronization of multi-source data in the time dimension; collected data should be stored in a unified structured format, including field names, units, and verification information, providing standardized support for subsequent data parsing, transmission, and processing.

[0034] Sensor accuracy is standardized: sensor accuracy parameters: weighing sensor accuracy ≤ ±0.1kg, flow meter accuracy ≤ ±0.5% FS (full scale), density meter accuracy ≤ ±0.005g / cm³, and electricity meter accuracy ≤ ±0.2 grade; a clear sensor calibration cycle is defined (calibrated once every 15 days, using the national metrological standard reference material comparison method) to avoid subjective errors caused by differences in sensor accuracy and ensure consistent data acquisition accuracy when reproducing different projects.

[0035] Sampling frequency and synchronization mechanism quantification: Unify the sampling frequency of all data (the original handover only mentioned "meeting the minimum requirements"): power consumption data once per minute, material addition data once per 5 minutes, mud characteristics (density, viscosity) data once per 10 minutes; time synchronization adopts GPS timing module (error ≤ 1ms), replacing the original vague expression of "time stamp alignment", ensuring that the time dimension of multi-source data is completely consistent, and improving data comparability and reproducibility.

[0036] The cost calculation in S3 above specifically includes: S3-1, Definition of Cost Items: Classify and standardize various cost elements involved in cost accounting, and clarify the accounting scope and attribution standards; Cost items are mainly divided into two categories: one is direct costs, including raw material consumption (such as bentonite and soda ash), energy expenditure (such as electricity costs), and on-site labor costs, which are directly related to the actual input of resources during construction; the other is indirect costs, including equipment depreciation, daily maintenance costs, and potential environmental penalties, which are expenditures that cannot be directly attributed to a certain operation link; S3-2, Baseline Cost Calculation: Based on the collected information on raw material usage, unit price, labor costs, and equipment depreciation, a standardized baseline cost calculation model is established. By summing the products of various raw material usages and their unit prices, adding the product of labor rates and working hours, and factoring in equipment depreciation costs, the basic cost value for the construction period is calculated. The calculation formula is as follows: ; Among them, Q i Let P be the amount of the i-th raw material used. i Let L be the unit price of the raw material, L be the labor cost rate, T be the labor hours, and D be the labor cost. equip Equipment depreciation costs; Equipment depreciation cost D equip The calculation introduces the "equipment operating load correction factor" δ: δ = actual operating time / rated operating time, and the corrected D equip = (Original value of equipment / Estimated useful life / 365) × δ, to ensure that the equipment depreciation cost matches the actual usage intensity and avoid accounting deviations caused by excessive depreciation of idle equipment; Automatic verification of raw material unit price: The "Market Price Database Integration Module" has been added: After the raw material unit price is manually entered, the system automatically connects to the real-time price platform of the industry (such as the shield tunneling mud material section of "China Building Materials Network"). If the entered unit price deviates from the average price of the platform by more than ±10%, a "Unit Price Anomaly Alert" will be triggered, and the average price of the platform and the deviation value will be displayed. The system requires secondary confirmation by the management personnel before it can be entered, thus avoiding subjective errors or intentional deviations in manual entry.

[0037] Electronic traceability of work hours records: replacing the original "manual record-keeping of work hours by work teams": adopting the method of "facial recognition check-in + equipment operation log association", work hours = equipment running time (electricity meter data) × number of on-site personnel (check-in data). For example: if the slurry pump runs for 8 hours and 2 people check in on-site, then the work hours = 8 × 2 = 16 person-hours. This avoids the subjective difference of manual estimation of work hours, and the work hours records can be traced back to specific equipment and personnel.

[0038] S3-3, Dynamic Correction Factor: A dynamic adjustment mechanism is introduced based on the baseline cost calculation to improve the timeliness and adaptability of cost accounting. By introducing the historical deviation rate α and the real-time operating condition coefficient β, the baseline cost is dynamically corrected, and the predicted cost C is calculated. pred The corrected formula is as follows: ; Among them, α is derived from the regression analysis results of historical data and reflects the trend of statistical deviation; β is the working condition correction factor that reflects the current changes in mud characteristics and is used to dynamically adjust the cost impact caused by changes in physical states such as sand content and viscosity. S3-4, Risk Assessment: The predicted cost is calculated multiple times using Monte Carlo simulation to assess the probability of exceeding the budget, thereby quantifying cost risks during construction. The system is based on dynamically corrected predicted costs. The sampling is repeated within a set number of simulations N, and it is determined whether the result exceeds the set budget value B each time; finally, the probability of cost overrun P is calculated using the following formula. over : ; Where I(⋅) is an indicator function used to determine whether the current simulation result exceeds the budget.

[0039] The sensors configured in S1 above have a one-to-one correspondence with the collected data: the weighing sensor is used to collect data on the amount of bentonite, slurry agent, thickener, and filter aid added, in kg; the flow meter is used to collect data on the flow rate of slag, clean water, and fresh slurry, in m³ / h; the densitometer is used to collect data on the density of the mud, in g / cm³; the electricity meter is used to collect data on the power consumption of various pumps, motors, and hydrocyclones, in kW·h; and the PLC monitoring device is used to collect data on the amount of flocculant added to the filter press and centrifuge, in ml.

[0040] The specific process of outlier identification in isolated forests is as follows: anomaly score is calculated for each numerical data sample based on the isolated forest algorithm. An anomaly score threshold of 0.6 is set. When the anomaly score of a sample is greater than 0.6, the sample is determined to be an outlier and is directly removed. The specific process of dynamic time warping is as follows: the sampling time of the power consumption data of the pre-screening unit is used as the reference time axis. The time series data of the mud-water separation unit, the slurry preparation unit, the pulping unit, and the waste slurry treatment unit are adjusted by the dynamic time warping algorithm so that the timestamp deviation of all time series data does not exceed 1 minute.

[0041] Quantification and Validation of the Anomaly Threshold in Isolation Forest: Clarify the core parameters of the Isolation Forest algorithm (the original disclosure only mentioned a threshold of 0.6): the number of decision trees is 100, and the sample subsample size is 256; prove the rationality of the threshold through "historical anomaly data backtesting" (e.g., select 1000 sets of known anomaly data (such as jump values ​​caused by sensor failure), and the recognition accuracy of using a threshold of 0.6 is ≥95%), and supplement the anomaly data alternative scheme (after removing outliers, use the moving average of 5 adjacent normal data to fill in the gaps, avoiding subjective selection of the filling method).

[0042] The specific equipment and data collection requirements for the above-mentioned power consumption data are as follows: power of the pre-screening unit, pump motor power of the first-stage hydrocyclone unit, vibration power of the dewatering screen unit, pump motor power of the second-stage hydrocyclone unit, power of the clear water pump motor in the clear water tank, power of the slurry mixing pump motor and stirring pump motor in the slurry mixing tank, power of the return slurry pump motor in the return slurry tank, power of the waste slurry pump motor in the waste slurry tank, power of the filter press pump motor in the waiting-to-press tank, power of the stirring pump motor, filtrate water pump motor and main motor power in the filter press, and power of the centrifuge, centrifuge power, two flocculant pump motors, and clear water pump motor power in the centrifuge; the sampling frequency for all equipment power consumption data shall not be less than once per minute.

[0043] The specific materials and collection methods corresponding to the above-mentioned material consumption data are as follows: the sludge removal volume of the sedimentation tank is collected manually, in m³; the pulping agent of the pulping system is collected by weighing equipment, in kg; the bentonite in the new pulp tank is collected by weighing equipment, and the feed rate of the new pulp pump and the addition rate of the shear pump are collected by flow meters, in kg and m³ / h respectively; the thickener in the slurry conditioning tank and the return slurry tank is collected by weighing equipment, in kg; the filter aid in the waste slurry tank and the waiting press tank is collected by weighing equipment, in kg; the flocculant addition rate of the filter press is recorded by the PLC system, in ml; and the flocculant addition rate and the clean water addition rate of the centrifuge are recorded by the PLC system, in ml.

[0044] The above historical deviation rate The calculation process is as follows: Select historical data from the last three construction cycles, calculate the difference between "historical actual cost - historical benchmark cost" in each cycle, and then take the arithmetic mean of the ratio of the difference to the historical benchmark cost of the corresponding cycle to obtain the historical deviation rate. If the historical actual cost is greater than the historical benchmark cost, Take a positive value; if the historical actual cost is less than the historical benchmark cost, Take the negative value; The calculation precision is retained to 4 decimal places.

[0045] The above-mentioned working condition coefficients The determination is based on the standard sand content of the mud. (Value taken as 5%), standard viscosity (Taking 25s as the benchmark) when the actual sand content Each higher At 1%, Increase by 0.02 from 1.0; when the actual sand content... Every lower At 1%, Reduce by 0.01 from 1.0; when the actual viscosity Each higher At 2s, Increase by 0.03 from 1.0; when the actual viscosity Every lower At 2s, Reduced by 0.015 from 1.0; The calculation result is rounded to three decimal places.

[0046] The above Monte Carlo simulation sets the number of simulations. No less than 1000 times; for each sampling, the amount of raw materials used... Labor cost rate Equipment depreciation costs Applying ±5% random fluctuations, based on collected actual data, the sampling process follows a normal distribution; after the simulation, if the probability of cost overrun is... If the percentage exceeds 20%, a cost warning will be triggered.

[0047] The preset time periods in S4 include daily, weekly, and monthly periods, and user-defined periods are also supported, with a period length range of 1-30 days. When classifying and collecting costs, material costs are subdivided into "bentonite cost, slurry agent cost, thickener cost, filter aid cost, and flocculant cost", electricity costs are subdivided into "electricity costs of equipment in each process", and spare parts costs are subdivided into "cost of vulnerable parts (screens, filter plates) and cost of core components (motors, pump bodies)", ensuring that the collection error of each subdivided cost does not exceed 2%.

[0048] The aforementioned preset thresholds are set as follows: based on historical cost data from the past 6 months, the historical average value of each cost item is calculated. with standard deviation The preset threshold range is When the actual value of a certain cost item exceeds the specified range, the system automatically marks it as an "abnormal cost item" and notes the corresponding process, equipment, and data acquisition time in the cost analysis report.

[0049] The application process of the aforementioned auxiliary analysis indicators is as follows: Correlation analysis is performed between mud density and viscosity data and raw material usage data. When the mud density is below 1.1 g / cm³, a suggestion to "increase the amount of bentonite added" is automatically generated; when the mud viscosity is above 30s, a suggestion to "reduce the amount of thickener added" is automatically generated. The suggestions include specific adjustment amounts: a "characteristic-usage" mapping formula fitted based on historical data. ,in To adjust the coefficient, These are actual characteristic values. The standard characteristic values ​​are used, and the cost prediction model is updated synchronously.

[0050] Example 2: A calculation method for the operation and control of mud treatment in slurry balance shield tunneling machines includes the following steps: Step S1: Data Acquisition; Weighing sensors, flow meters, density meters, electricity meters, and PLC monitoring devices are installed on material conveying devices, liquid storage tanks, pumps, and chemical dosing devices at key process nodes such as the pre-screening unit, slurry separation unit, slurry preparation unit, slurry treatment unit, and waste slurry treatment unit of the shield tunneling construction system. These devices are used to collect real-time information on the addition amount and flow rate of materials such as bentonite, slurry preparation agent, thickener, filter aid, and slag, as well as the power consumption data of various pumps, motors, and hydrocyclones. Through system integration, ledger management, and manual input, relevant data on equipment depreciation are collected, including the original value of the equipment, its service life, current usage time, or remaining usable years. Simultaneously, an operating cost accounting model is obtained. The steps for constructing the slurry treatment operating cost model for slurry balance shield tunneling are as follows: Step S1-1: Define Data Items; Classify and define the data types required by the cost accounting application system, and clarify their collection methods and units, serving as the basis for subsequent data collection and analysis. Data items mainly include four categories: First, equipment power consumption data, primarily collecting the power consumption of various operating equipment, collected via electricity meters, with units of kW·h; Second, material consumption data, involving the usage of raw materials such as bentonite, soda ash, thickeners, and filter aids, collected through weighing equipment, PLC systems, or manual recording, with units including kg or m³; Third, auxiliary analysis index data, including mud characteristic data and equipment operating status data, used to support cost analysis and operational efficiency assessment; Fourth, manually entered cost data, mainly including economic parameters such as raw material unit prices, labor rates, and equipment depreciation rates, which are regularly maintained and entered by management personnel. S1-1 also includes auxiliary analysis indicators for equipment operating temperature data: Temperature sensors are installed on the casings of pumps and motors, with a measurement range of -20℃ to 150℃ and an accuracy of ±0.5℃, to collect real-time equipment temperature. When the temperature exceeds 80℃, the system automatically correlates this with power consumption data to calculate the "temperature-power loss coefficient" γ, where γ = 1 + 0.01 × (T 实际 -80), T 实际 The real-time temperature of the equipment is calculated, and γ is incorporated into the cost accounting model to adjust the equipment's electricity costs. Step S1-2: Define data standards; standardize the accuracy requirements, time synchronization methods, and data structures of the collected data; sensor data must meet the set measurement accuracy standards to avoid interference from errors affecting subsequent calculation results; all data should be aligned by timestamps, and the sampling frequency should meet the minimum acquisition requirements to ensure consistent synchronization of multi-source data in the time dimension; the collected data should be stored in a unified structured format, including field names, units, and verification information, to provide standardized support for subsequent data parsing, transmission, and processing; Step S2: Data Preprocessing; For the collected multi-source heterogeneous data, wavelet transform is used to denoise the text, and breadth-first search is combined to extract key indicators, achieving the required structure and standardization; Simultaneously, Z-score standardization, isolated forest outlier identification, and dynamic time warping are performed on numerical data to unify the data format, remove outlier interference, and align different time series data to ensure data consistency and usability; The data preprocessing for slurry treatment in slurry balance shield tunneling includes the following steps: Step S2-1: Data Preprocessing. Multi-level preprocessing is performed on the collected raw data and requirement information to ensure consistency and usability in subsequent analysis. First, for requirement data from text input, wavelet transform is used for noise reduction to remove semantic interference and noise content. A breadth-first search algorithm is then used to extract key indicator parameters, achieving structured and standardized requirement information. Next, the collected numerical raw data is normalized using Z-score standardization to eliminate dimensional differences and improve data comparability. To identify and remove outliers, the Isolation Forest algorithm is introduced. Anomaly scores are calculated based on path length; values ​​exceeding a set threshold (e.g., 0.6) are considered outliers and removed. Finally, for data sources with inconsistent sampling frequencies or different time bases, Dynamic Time Warping (DTW) is used for time series alignment to minimize deviations in various data types along the time axis, ensuring the use of all data within a unified time dimension.

[0051] Step S3, Cost Calculation: Based on the high-quality data collected and preprocessed, a standard cost model is constructed according to raw material usage, electricity consumption, labor costs, and equipment depreciation. Combined with energy consumption change curves at each stage, the dynamic loss of sub-items is further quantified by calculating the energy consumption difference and energy consumption ratio between adjacent levels of different processes, forming a time-series-oriented cost loss curve. Historical deviation correction factors and operating condition coefficients are introduced to dynamically adjust costs, and the cost accounting results for each time period and each functional module are comprehensively output. The cost calculation for slurry treatment in slurry balance shield tunneling includes the following steps: Step S3-1: Cost Item Definition; Classify and standardize the various cost elements involved in cost accounting, and clarify the accounting scope and attribution standards. Cost items are mainly divided into two categories: one is direct costs, including raw material consumption (such as bentonite and soda ash), energy expenditure (such as electricity costs), and on-site labor costs, which are directly related to the actual input of resources during construction; the other is indirect costs, including equipment depreciation, daily maintenance costs, and potential environmental penalties, which are expenditures that cannot be directly attributed to a certain operation. Step S3-2: Baseline Cost Calculation; Based on the collected information on raw material usage, unit price, labor costs, and equipment depreciation, a standardized baseline cost calculation model is established. By summing the products of the usage and unit price of various raw materials, adding the product of labor rate and working hours, and factoring in equipment depreciation costs, the basic cost value for the construction period is calculated. The calculation formula is as follows: ; Among them, Q i Let P be the amount of the i-th raw material used. i Let L be the unit price of the raw material, L be the labor cost rate, T be the labor hours, and D be the labor cost. equip Equipment depreciation cost; Equipment depreciation cost D equip The calculation introduces an "equipment operating load correction factor" δ: δ = actual operating time / rated operating time, with the rated operating time calculated as 8 hours / day. After correction, Dequip = (original equipment value / estimated service life / 365) × δ, ensuring that the equipment depreciation cost matches the actual usage intensity and avoiding accounting deviations caused by excessive depreciation of idle equipment. Step S3-3: Dynamic Adjustment Factor; A dynamic adjustment mechanism is introduced based on the baseline cost calculation to improve the timeliness and adaptability of cost accounting. By introducing the historical deviation rate α and the real-time operating condition coefficient β, the baseline cost is dynamically adjusted, and the predicted cost C is calculated. pred The corrected formula is as follows: ; Among them, α is derived from the regression analysis results of historical data and reflects the trend of statistical deviation; β is the working condition correction factor that reflects the current changes in mud characteristics and is used to dynamically adjust the cost impact caused by changes in physical states such as sand content and viscosity. Step S3-4: Risk Assessment; The predicted cost is calculated multiple times using Monte Carlo simulation to assess the probability of exceeding the budget, thereby quantifying the cost risk during construction. The system is based on dynamically corrected predicted costs. The sampling is repeated within a set number of simulations N, and it is determined whether the result exceeds the set budget value B each time. Finally, the probability of cost overrun P is calculated using the following formula. over : ; Where I(⋅) is an indicator function used to determine whether the current simulation result exceeds the budget; Step S4, Total Cost Summary: Step S5, Total Cost Summary; Based on the completed individual cost calculation results, the system classifies and collects material costs, electricity costs, spare parts costs, labor costs, equipment depreciation, etc., according to preset time periods (such as daily, weekly, monthly) and accumulates the values ​​to form a unified total cost data table, and performs structured summary of various cost factors. Step S5: Cost Analysis and Feedback; After the system completes the total cost summary, it generates a cost analysis report according to the set period, performs statistical analysis and trend judgment on various cost data, and identifies cost items with abnormal fluctuations or exceeding the preset threshold; at the same time, it combines the operation data of each process to locate potential high consumption points or efficiency deviations, so as to realize continuous monitoring and feedback adjustment of the cost structure.

[0052] In this embodiment, the collected data items mainly include four categories: first, equipment power consumption data; second, material consumption data; third, auxiliary analysis indicator data; and fourth, manually entered cost data. Furthermore, equipment power consumption data mainly collects the electricity consumption of various operating equipment, collected via electricity meters, with the unit being kW·h; material consumption data involves the usage of raw materials such as bentonite, soda ash, thickeners, and filter aids, collected through weighing equipment, PLC systems, or manual recording, with units including kg or m³; auxiliary analysis data includes mud characteristic data and equipment operating status data, used to support cost analysis and operational efficiency assessment; manually entered cost data mainly includes economic parameters such as raw material unit price, labor rate, and equipment depreciation rate, which are regularly maintained and entered by management personnel. Specifically, the data items required in the cost accounting application system are categorized as follows: Equipment power consumption data:

[0053] Material consumption data:

[0054] Auxiliary analysis indicators and data:

[0055] Spare parts replacement record: Sources: Spare parts requirement form, inventory receipt form, usage form (manual form); Typical spare parts: screening machine vibration motor, pump, filter plate, hydrocyclone liner, etc.; Labor costs: Sources of records: work hours of work teams, construction plans, and standard labor unit prices; Equipment depreciation: Calculation basis: original value of equipment, service life, and amortization period (days / months).

Claims

1. A method of slurry balance shield tunneling slurry handling operation control calculation, characterized in that the method Comprise: S1, data acquisition; in the pre-screening unit, slurry separation unit, mud unit, mud unit, waste mud treatment unit process node of shield construction system, material conveying device, liquid storage tank, pump and dosing device are respectively configured with weighing sensor, flow meter, densimeter, ammeter and PLC monitoring device, for real-time acquisition of bentonite, mud preparation agent, tackifier, filter aid, slag material addition amount and flow information, and power data of various pumps, motors and cyclones; through system docking, account management and manual input of depreciation related data of collection equipment, including equipment original value, service life, current use time or remaining available life; at the same time, the model of operation cost accounting is obtained; S1 also includes sensor calibration and synchronous control: every set number of days, all sensors are calibrated by national measurement standard substance comparison method, wherein the weighing sensor calibration error needs to be less than or equal to ± 0.1 kg, the flow meter calibration error is less than or equal to ± 0.5% FS under full range condition, the densimeter calibration error is less than or equal to ± 0.005 g / cm3, and the ammeter calibration error is less than or equal to ± 0.2 level; the time synchronization of all collection equipment is realized through GPS time service module, the time synchronization error is less than or equal to 1 ms, and the consistency of multi-source data in time dimension is ensured; S2, data preprocessing; S3, cost calculation; S4, total cost summary; according to the completed calculation results of various single costs, the system classifies and collects the material cost, electricity cost, spare part cost, labor cost, equipment depreciation and the like according to the preset time period, and accumulates the numerical values to form a unified total cost data table, and the various cost factors are structurally summarized; S5, cost analysis and feedback; the continuous monitoring and feedback adjustment of cost structure are realized.

2. The method according to claim 1, wherein, The specific process of S2 is: The data collected in S1 is multi-source heterogeneous data, for multi-source heterogeneous data, wavelet transform method is used for noise reduction processing of text, and key indicators are extracted by combining breadth-first search to realize the structuring and standardization of demand; at the same time, Z-score standardization, isolated forest abnormal value identification and dynamic time warping processing are performed on numerical data, the data format is unified, abnormal interference is eliminated, and different time sequence data is aligned to ensure data consistency and availability.

3. The method according to claim 2, wherein, The specific process of S5 is: After the system completes the total cost summary, the cost analysis report is generated according to the set period, the statistical analysis and trend judgment of various cost data are carried out, the cost items with abnormal fluctuations or exceeding the preset threshold are identified; at the same time, combined with the running data of each process, the potential high consumption point or efficiency deviation is located, the continuous monitoring and feedback adjustment of cost structure are realized; S5 also includes mud property-cost linkage feedback: according to the correlation model of cost analysis results and mud density, viscosity data, when the mud density is lower than p 标准 and the bentonite cost is not over limit, automatically generate a control instruction to "increase the bentonite addition amount", which contains the specific adjustment amount, based on the "density-bentonite consumption" mapping formula fitted based on historical data: AQ=k x (p 标准 -p 实际 ), where k is the adjustment coefficient, the value range is 1.2-1.5 kg / (g·cm -3 ), p 标准 =1.1 g / cm³; when the mud viscosity is higher than 30 s and the thickener cost is over limit, automatically generate a composite instruction to "reduce the thickener addition amount + adjust the stirring pump speed", the stirring pump speed adjustment range is positively correlated with the viscosity over-limit value, and the working condition coefficient β in the cost accounting model is updated synchronously.

4. The method of claim 1, wherein, In the process of obtaining the model of operation cost accounting in S1, it also includes: S1-1, Collect data item definition; classify and define the data types required in cost accounting, and clearly define the collection method and unit, which is the basis for subsequent data collection and analysis; Data items mainly include four categories: First, equipment power consumption data, collect the power consumption of various running equipment, the collection method is electric meter, the unit is kW·h; Second, material consumption data, involving the use of bentonite, soda ash, tackifier, filter aid raw materials, collected by weighing equipment, PLC system or manual recording method, units including kg or m³; Third, auxiliary analysis indicators, including mud property data and equipment operating state data, used to support cost analysis and operating efficiency evaluation; Fourth, manually entered cost data, including raw material unit price, labor rate and equipment depreciation rate economic parameters, maintained and entered by management personnel on a regular basis; The auxiliary analysis index also includes device operating temperature data: by installing temperature sensors on the pump and motor device shell, real-time temperature of the device is collected, when the temperature exceeds 80°C, the power consumption data is automatically associated to calculate the "temperature-power consumption loss coefficient" γ, γ = 1 + 0.01 × (T 实际 -80), T 实际 is the real-time temperature of the device, and γ is included in the cost accounting model to correct the device power consumption cost; S1-2, Data standard definition; unify and standardize the precision requirements, time synchronization methods and data structures of collected data; Sensor data must meet the set measurement accuracy standard to avoid error interference affecting subsequent accounting results; All data are aligned by timestamp, and the sampling frequency meets the minimum collection requirement to ensure consistent synchronization of multi-source data in the time dimension; Collected data are stored in a unified structured format, including field name, unit and verification information, providing standardized support for subsequent data parsing, transmission and processing.

5. The method of claim 2, wherein the method further comprises: The cost calculation in S3 specifically includes: S3-1, Cost item definition; classify and standardize various cost elements involved in cost accounting, and clearly define the accounting scope and attribution standard; Cost items are divided into two categories: First, direct costs, including raw material consumption, energy expenditure and on-site labor costs, which are directly related to the actual input of resources in the construction process; Second, indirect costs, including equipment depreciation, routine maintenance costs and potential environmental penalties, which are not directly attributed to a work link; S3-2, Reference cost calculation; based on the collected raw material usage, unit price, labor cost and equipment depreciation information, a standardized reference cost calculation model is established; By multiplying the usage of each type of raw material by its unit price, adding the product of labor rate and working hours, and adding equipment depreciation cost, the basic cost value in the construction period is calculated; The calculation formula is: ； wherein Q i is the amount of the i-th raw material used, P i is the unit price of the raw material, L is the labor rate, T is the man-hours, D equip is the depreciation cost of the equipment; D = (P - S) / n equip Calculate the "equipment running load correction factor" δ: δ = actual running time / rated running time, and correct D equip = (P - S) / n × δ, ensure that the equipment depreciation cost matches the actual use intensity, avoid accounting deviation caused by idle equipment over-depreciation; S3-3, dynamic correction factor; introduce dynamic adjustment mechanism on the basis of benchmark cost calculation, in order to improve the timeliness and adaptability of cost accounting; by introducing historical deviation rate α and real-time working condition coefficient β, the benchmark cost is dynamically corrected, and the predicted cost C is calculated pred ; The correction formula is as follows: ； Where, α is obtained from the regression analysis results of historical data, reflecting the statistical bias trend; β is a working condition correction factor reflecting the change of mud properties, used to dynamically adjust the cost impact of changes in physical state such as sand content and viscosity; S3-4, risk assessment; through Monte Carlo simulation method, the predicted cost is simulated many times, the probability of exceeding the budget is evaluated, and the cost risk in the construction process is quantified; the system is based on the dynamically corrected predicted cost , the sampling is repeated within the set simulation times N, and it is judged whether the result of each time exceeds the set budget value B; finally, the cost overrun probability P is calculated by the following formula over : ； Where, I(⋅) is an indicator function, used to determine whether the current simulation result exceeds the budget.

6. The method of claim 1, wherein, The sensor configured in S1 has a one-to-one correspondence with the collected data: the weighing sensor is used to collect the data of the added amount of bentonite, pulping agent, tackifier and filter aid, with the unit being kg; the flow meter is used to collect the flow data of the residue, clean water and new slurry, with the unit being m³ / h; the density meter is used to collect the density data of the slurry, with the unit being g / cm³; the electric meter is used to collect the power consumption data of various pumps, motors and cyclones, with the unit being kW·h; the PLC monitoring device is used to collect the flocculant addition data of the filter press and centrifuge, with the unit being ml.

7. The method of claim 2, wherein the method further comprises: The specific process of the isolated forest anomaly value identification is that the anomaly score of each numerical data sample is calculated based on the isolated forest algorithm, the anomaly score threshold is set to 0.6, and when the sample anomaly score is greater than 0.6, the sample is determined to be an outlier and is directly excluded; the specific process of the dynamic time warping processing is that the power consumption data sampling time of the pre-screening unit is taken as the reference time axis, the time series data of the slurry separation unit, the mixing unit, the pulping unit and the waste slurry treatment unit are adjusted through the dynamic time warping algorithm, so that the timestamp deviation of all time series data does not exceed 1 minute.

8. The method of claim 4, wherein the method further comprises: The specific equipment corresponding to the equipment power consumption data and the collection requirements are as follows: the pre-screening unit collects the power of the screening machine, the first-stage cyclone unit collects the power of the pump motor, the dewatering screen unit collects the vibration power, the second-stage cyclone unit collects the power of the pump motor, the clean water tank collects the power of the clean water pump motor, the mixing tank collects the power of the mixing pump motor and the stirring pump motor, the back slurry tank collects the power of the back slurry pump motor, the waste slurry tank collects the power of the waste slurry pump motor, the pressure tank collects the power of the filter material pump motor, the filter press collects the power of the stirring pump motor, the filtrate water pump motor and the main motor, and the centrifuge collects the power of the centrifuge, the power of the two flocculant pump motors and the power of the clean water pump motor; the sampling frequency of all equipment power consumption data is not less than 1 time / minute.

9. The method of claim 4, wherein the method further comprises: The specific materials corresponding to the material consumption data and the collection methods are as follows: the amount of mud removed from the sedimentation tank is collected by manual recording, with the unit being m³; the pulping agent of the pulping system is collected by the weighing device, with the unit being kg; the bentonite of the new slurry tank is collected by the weighing device, the new slurry pump feeding amount and the shear pump addition amount are collected by the flow meter, with the units being kg and m³ / h respectively; the tackifier of the mixing tank and the back slurry tank is collected by the weighing device, with the unit being kg; the filter aid of the waste slurry tank and the pressure tank is collected by the weighing device, with the unit being kg; the addition amount of the flocculant of the filter press is recorded by the PLC system, with the unit being ml; the addition amount of the flocculant and the addition amount of the clean water of the centrifuge are recorded by the PLC system, with the unit being ml.

10. The method of claim 5, wherein, The historical deviation rate The calculation process of the historical deviation rate is as follows: selecting historical data of the last three construction periods, calculating the difference between the historical actual cost and the historical benchmark cost in each period, and then taking the arithmetic average of the ratio of the difference to the historical benchmark cost of the corresponding period to obtain the historical deviation rate ; if the historical actual cost is greater than the historical benchmark cost, a positive value is taken; if the historical actual cost is less than the historical benchmark cost, a negative value is taken; the calculation precision is retained to 4 decimal places.

11. The method of claim 5, wherein the method is characterized by: The working condition coefficient The determination basis is: taking the standard sand content of mud , standard viscosity as the basis, when the actual sand content is higher than 1%, 0.02 is added to 1.0; when the actual sand content is lower than 1%, 0.01 is reduced from 1.0; when the actual viscosity is higher than 2s, 0.03 is added to 1.0; when the actual viscosity is lower than 2s, 0.015 is reduced from 1.0; The calculation result of is kept to 3 digits after the decimal point.

12. The method of claim 5, wherein the method is characterized by: The setting simulation times of the Monte Carlo simulation No less than 1000 times; for raw material usage , artificial fee rate , equipment depreciation cost , respectively apply ±5% random fluctuation, sampling process meets normal distribution; after simulation is completed, if cost overrun probability is greater than 20%, cost early warning is triggered.

13. The method of claim 1, wherein the method is characterized by: The preset time period in S4 includes daily, weekly and monthly periods, and supports user-defined periods; during classification and collection, the material cost is subdivided into "bentonite cost, pulping agent cost, tackifier cost, filter aid cost and flocculant cost", the electricity cost is subdivided into "electricity cost of devices in each process", and the spare part cost is subdivided into "cost of consumable parts and cost of core components", so as to ensure that the collection error of each type of subdivided cost does not exceed 2%.

14. The method of claim 3, wherein the method is characterized by: The preset threshold is set based on historical cost data of the past 6 months to calculate the historical average value of each cost item and standard deviation The preset threshold range is When the actual value of a certain cost item exceeds the range, the system automatically marks it as an "abnormal cost item" and notes the corresponding process, equipment, and data collection time in the cost analysis report.

15. The method of claim 4, wherein the method is characterized by: The application process of the auxiliary analysis index is: correlating and analyzing the mud density, viscosity data and raw material dosage data, automatically generating the suggestion of "increasing the bentonite addition amount" when the mud density is lower than 1.1 g / cm3, automatically generating the suggestion of "reducing the tackifier addition amount" when the mud viscosity is higher than 30 s, and the suggestion contains specific adjustment amount: the "property-dosage" mapping formula based on historical data fitting: wherein is an adjustment coefficient, is an actual property value, is a standard property value, and the cost prediction model is updated synchronously.

Images