Heat pump energy optimization operation system and method applied to farm

Abstract

The invention discloses a heat pump energy optimization operation system and method applied to a farm, and relates to the technical field of control optimization. Different sensors are used for collecting working data when a heat pump energy application system works; collecting physical data of the heat pump energy application system, and constructing a digital twinborn model of the heat pump energy application system according to the physical data and the working data; constructing a prediction model by using a machine learning algorithm, and predicting future thermal load and waste heat through the prediction model; calculating a comprehensive energy efficiency ratio of the heat pump energy application system by using the working data, setting constraint conditions, and solving an optimal operation variable of the heat pump energy application system by using a model prediction control algorithm; a control instruction is generated through the optimal operation variable, and the operation state of the heat pump energy application system is adjusted through the control instruction; and a prediction value and a true value in the prediction model during real-time operation of the heat pump energy application system are collected, and a prediction error optimization model is calculated.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of control optimization, in particular to a heat pump energy optimization operation system and method applied to a farm. BACKGROUND

[0002] Early farms mostly use longitudinal ventilation mode of fan ventilation + water curtain cooling + boiler heating. However, the air flow in the livestock and poultry house is complex, and the three systems are difficult to adapt to each other, which is prone to cause large temperature difference between front and back and between top and bottom in the house. Such temperature difference will cause cold and heat stress reaction of the breeding animals, and further induce diseases and affect growth efficiency. Moreover, the paper water curtain is prone to harden due to water quality problem in long-term use, which not only reduces the cooling and ventilation effect, but also may cause insufficient oxygen supply in the house. In greenhouse breeding, the temperature of fresh air in winter is too low, and the energy consumption of the simple heating device is also very large. Traditional boilers mostly rely on coal and gas, and a large amount of pollutants will be discharged during combustion. With the tightening of global environmental protection policies, the use of such boilers is strictly limited, and the breeding farm has to face the additional cost of equipment modification or replacement. At the same time, the prices of fossil energy such as coal and fuel oil fluctuate greatly, and long-term use will make the energy cost of breeding high. For example, when some large-scale chicken farms use fuel oil boilers, the fuel oil cost is high in winter, and the cost pressure is significant. With the gradual development of the breeding industry towards scale and intensification, the breeding density is continuously improved. Under the condition of high-density breeding, the animals and manure will generate a large amount of heat and harmful gases such as ammonia, which not only requires higher temperature control stability, but also needs to consider ventilation and air purification. Traditional temperature control equipment is difficult to meet the comprehensive demand of temperature control, ventilation and low consumption, while the heat pump system can combine heat exchange, fresh air treatment and other functions to adapt to the complex scene of intensive breeding.

[0003] However, in the traditional system, it is assumed that the temperature (35℃) and flow of biogas slurry are stable, but the actual production of the farm is fluctuating. For example, the temperature of biogas slurry may change with the season, the amount of flushing water and the fermentation efficiency; the heat load of the fermentation tank may also change due to the amount of feed and environmental temperature. Dynamic fluctuations will cause the heat pump system to run under fixed conditions, either the boiler is frequently started and stopped due to insufficient energy supply, or the waste heat cannot be fully recovered, and the system comprehensive energy efficiency (SCOP) decreases; and the traditional heat pump energy optimization device involves a wide range of sites, and it is difficult to analyze and optimize the data of the device. SUMMARY

[0004] The purpose of the present application is to provide a heat pump energy optimization operation system and method applied to a farm to solve the problems in the prior art.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical scheme: A heat pump energy optimization operation method applied to a farm, the method comprising the following steps: S100, constructing a heat pump energy application system in a farm, and collecting working data by using different sensors when the heat pump energy application system is working; Further, the specific steps of collecting heat source, load, environment and device running data as working data by using different sensors when the heat pump energy application system is working are as follows: S101, constructing a heat pump energy application system in a farm, wherein the heat pump energy application system comprises a biogas slurry pool, a heat exchanger, a heat pump, a boiler and a fermentation tank, the heat pump comprises an evaporator and a condenser, the biogas slurry pool transmits waste heat to the heat exchanger, the heat exchanger generates hot water by using the waste heat, the hot water enters the heat pump, the heat pump re-heats the hot water and inputs the hot water into the fermentation tank, the boiler supplies heat to the heat pump when the heat pump re-heats the hot water, and all devices are connected by pipelines; S102, collecting working data by using different sensors when the heat pump energy application system is working, wherein the working data comprises biogas slurry temperature, biogas slurry flow, fermentation tank temperature, environment temperature, heat pump power consumption and boiler heat supply amount; and the sensors transmit the collected data to a central processor through the Internet of Things.

[0006] The full-dimensional data of heat source, load, environment and device running are covered to provide accurate data sources for subsequent model construction and optimization. The data is transmitted by the Internet of Things to ensure real-time data and support subsequent real-time regulation and control decisions.

[0007] S200, collecting physical data of the heat pump energy application system, and constructing a digital twin model of the heat pump energy application system according to the physical data and the working data; and constructing sub-models in the digital twin model, including a heat pump heating model, a heat exchanger model and a pipeline heat loss model; Further, the specific steps of constructing the digital twin model of the heat pump energy application system are as follows: S201, collecting physical data of the heat pump energy application system, obtaining a digital twin model of the heat pump energy application system by using the physical data, and constructing sub-models in the digital twin model, wherein the sub-models include a heat pump heating model, a heat exchanger model and a pipeline heat loss model; The three sub-models of heat pump heating, heat exchanger and pipeline heat loss are split to accurately reproduce the physical laws of device running and realize operation state visualization.

[0008] The construction method of the heat pump heating model is as follows: extracting a heat pump performance coefficient in the heat pump, constructing a heat pump heating model by using the heat pump performance coefficient and heat pump power consumption, and the formula is as follows: ; In the formula, Q heat represents heat pump heating capacity, COP represents heat pump performance coefficient, and W hp represents heat pump power consumption. The traditional system can only roughly determine the running state of the heat pump through the "output temperature", and cannot quantify the "matching relationship between energy consumption and heating capacity". This model can accurately identify the abnormal working condition of "COP drops but W hp rises", provide early warning of equipment failure, avoid energy waste caused by inefficient operation of equipment, and provide "quantitative basis" for subsequent optimization of heat pump operation variables.

[0009] The model parameters can be updated in real time to adapt to the "performance degradation" during long-term operation of the heat pump. Compared with the traditional fixed parameter model, it can more truly reflect the actual running state of the equipment and improve the simulation accuracy.

[0010] S202, the heat exchanger model construction method is: extracting the mass flow and constant pressure specific heat capacity of the heat exchanger, obtaining the heat exchange through the mass flow multiplied by the constant pressure specific heat capacity and then multiplied by the temperature difference, extracting the maximum and minimum heat exchange; using the minimum heat exchange to heat capacity flow rate ratio r of the maximum heat exchange, using the heat transfer unit method combined with the heat capacity flow rate ratio to calculate the heat exchanger efficiency, and calculating the actual heat exchange through the heat exchanger efficiency and the maximum heat exchange, the formula is: ; In the formula, Q sh represents the actual heat exchange, represents the heat exchanger efficiency, and Q max represents the maximum heat exchange; The traditional system only monitors the "inlet and outlet temperature difference" of the heat exchanger, and cannot accurately calculate the actual heat exchange, which may cause "insufficient waste heat recovery" or "excessive energy consumption"; This model calculates the maximum heat exchange through "mass flow x constant pressure specific heat capacity x temperature difference", and then optimizes the efficiency calculation combined with the heat capacity flow rate ratio, which can accurately locate the "energy efficiency bottleneck" of the heat exchanger: for example, when the r value is too large, it means that the heat capacity flow rate of the cold side and the hot side is not matched, which can be optimized by adjusting the water pump frequency to improve the flow and improve the heat exchanger efficiency, so that the actual heat exchange is close to the maximum heat exchange, the waste heat recovery efficiency of the biogas liquid is maximized, and the problems of "waste heat waste" and "inefficient energy consumption" are reduced.

[0011] S203, the construction method of the pipeline heat loss model is: collecting the surface area, heat transfer coefficient and water temperature of each connected pipeline in the heat pump energy application system, calculating the pipeline heat loss, the formula is: ; In the formula, Q loss represents the pipeline heat loss, U represents the heat transfer coefficient, A represents the pipeline surface area, T water represents the water temperature, and T amb represents the ambient temperature; the real-time calculation values of the three sub-models in the digital twin model are output as the running data of the heat pump energy application system.

[0012] Collect the "heat transfer coefficient, surface area, water temperature" and "environmental temperature" of each pipeline to realize the quantitative calculation of the heat loss of each pipeline: the traditional system often ignores the heat loss of the pipeline and only calculates the "heat pump output heat = heat obtained by the fermentation tank", which leads to the blind increase of the heat pump power consumption when the actual heat supply is insufficient. This model can accurately calculate the heat loss of each pipeline, provide a basis for subsequent "pipeline insulation optimization" and "heating capacity compensation", and further reduce energy waste; Integrate the real-time calculation values of heat pump heating, heat exchanger heat exchange, and pipeline heat loss into "virtual operation data", and form a "virtual-actual contrast" with the actual operation data of the physical device, which can intuitively find "abnormal points" in the physical system, realize "early fault identification and accurate problem positioning", and reduce the downtime loss of the traditional system "which can only be checked after the fault occurs".

[0013] The key parameters are calculated by the quantitative formula to provide a reliable virtual simulation basis for subsequent prediction and optimization.

[0014] S300, a prediction model is constructed by using a machine learning algorithm, historical working data of a heat pump energy application system is used to train the prediction model, and future heat load and waste heat are predicted by the prediction model; Further, the specific steps of predicting future heat load and waste heat by the prediction model are: S301, a heat load prediction model and a waste heat prediction model are respectively constructed by using a machine learning algorithm; for the heat load prediction model, the heat pump heating capacity, the boiler heat supplement capacity, and the pipeline heat loss in the history are extracted, the heat pump heating capacity plus the boiler heat supplement capacity minus the pipeline heat loss is used to obtain the heat load of the fermentation tank, and the historical production plan, the fermentation tank temperature, and the environmental temperature are extracted, the historical production plan, the fermentation tank temperature, the environmental temperature, and the heat load of the fermentation tank are used as training data, the heat load prediction model is trained, the prediction time is set to 1 hour, the real-time production plan, the fermentation tank temperature, and the environmental temperature are input into the heat load prediction model to predict the fermentation tank heat load in the next 1 hour; S302, for the waste heat prediction model, the historical biogas slurry flow, biogas slurry temperature, heat exchanger actual heat exchange capacity, and weather data are extracted, the biogas slurry flow, biogas slurry temperature, heat exchanger actual heat exchange capacity, and weather data are used as training data, the waste heat prediction model is trained, the prediction time is set to 1 hour, the real-time biogas slurry temperature, biogas slurry flow, and weather data are input into the waste heat prediction model to predict the heat exchanger actual heat exchange capacity in the next 1 hour, and the heat exchanger actual heat exchange capacity is used as the biogas slurry waste heat.

[0015] S400, the comprehensive energy efficiency ratio of the heat pump energy application system is calculated by using the working data, the constraint condition is set, and the optimal operation variable of the heat pump energy application system is solved by using the model predictive control algorithm; Furthermore, the specific steps for solving the optimal operating variables of the heat pump energy application system using the model predictive control algorithm are as follows: S401. Input the predicted heat load and predicted waste heat for the next hour into the digital twin model to extract the heat pump heating capacity, heat pump power consumption, and total system power consumption of the heat pump energy application system for the next hour. Calculate the comprehensive energy efficiency ratio of the heat pump energy application system using the following formula: ; In the formula, SCOP represents the overall energy efficiency ratio of the heat pump energy application system, W. hp This indicates the heat pump's power consumption, in W. pump This indicates the total system power consumption of a heat pump energy application system; Compared to the traditional single indicator of "only calculating the COP of heat pumps", SCOP covers the total power consumption of all devices such as heat pumps and water pumps, and can better reflect the overall energy utilization efficiency of the system. For example, under a certain operating condition, the COP of the heat pump is high, but the water pump has a large total power consumption due to excessive frequency. Traditional indicators will misjudge "the system has good energy efficiency", while SCOP can accurately identify the problem of "local high efficiency but overall low efficiency", providing a "global perspective" for optimization.

[0016] S402. In the digital twin model, under the standard of predicted heat load and predicted waste heat for the next hour, the heat pump condensing temperature, evaporating temperature, cold-side water pump frequency and hot-side water pump frequency in the heat pump energy application system are adjusted. After each adjustment, the corresponding comprehensive energy efficiency ratio is calculated, and the heat pump condensing temperature, evaporating temperature, cold-side water pump frequency and hot-side water pump frequency corresponding to the maximum comprehensive energy efficiency ratio are selected as the optimal operating variables. S403. When debugging the heat pump condensing temperature, evaporation temperature, cold-side water pump frequency and hot-side water pump frequency in the heat pump energy application system, set constraints. Specifically, in the digital twin model, based on professional experience, the staff sets the operating limits of the heat pump energy application system according to the physical laws of the three sub-models. The operating limits are the maximum values ​​that the fermenter temperature, operating pressure and system power can withstand when the heat pump energy application system is running.

[0017] With the goal of maximizing SCOP in the next hour, the four core operating variables of heat pump condensing temperature, evaporating temperature, cold-side water pump frequency, and hot-side water pump frequency were adjusted. Traditional control often adopts "trial and error with a single variable", which can easily lead to "local optimization but overall inefficiency". However, the model predictive control algorithm can consider the interaction of the four variables at the same time. Through multiple rounds of simulation and debugging, the optimal combination of "heating capacity meets demand and total power consumption is minimized" is selected.

[0018] The algorithm is highly efficient, completing a round of optimization calculations within 10-20 seconds. It is adapted to the control requirements of a "1-hour prediction cycle", ensuring that the optimal operating variables can be applied to actual operation in a timely manner and avoiding "optimization failure due to calculation lag".

[0019] S500: Generates control commands using optimal operating variables, and uses control commands to adjust the operating status of the heat pump energy application system; Furthermore, the specific steps for adjusting the operating status of the heat pump energy application system using control commands are as follows: S501: The real-time heat pump condensing temperature, evaporating temperature, cold-side water pump frequency, and hot-side water pump frequency in the heat pump energy application system are collected and compared with the optimal operating variables. The difference between the four operating variables is calculated, and the difference is converted into a control command and sent to the corresponding device for control. The control command is issued through the PLC. After the heat pump energy application system is adjusted by the control command of the optimal operating variables, it returns to the working data of the heat pump energy application system collected in S100, and then executes the subsequent steps to form a closed loop. S502. Extract the heat pump's heating capacity and the fermenter's heat load. When the heat pump's heating capacity is less than the fermenter's heat load, start the boiler to supplement heat. S600 collects predicted and actual values ​​from the prediction model during real-time operation of the heat pump energy application system, and calculates the prediction error to optimize the model.

[0020] Furthermore, the specific steps for calculating the prediction error optimization model are as follows: S601. In the operation of a heat pump energy application system, a model performance evaluation is performed weekly. Operating data and actual operating data are extracted from the digital twin model, and predicted values ​​and actual values ​​are extracted from the prediction model. These are then input into the error calculation model to calculate the errors of the digital twin model and the prediction model. The preset error calculation model is as follows: ; In the formula, Error represents the model error, measured represents the data value in the model, corresponding to the working data in the digital twin model and the predicted value in the prediction model, and predicted represents the actual measured value, corresponding to the actual working data and the actual value. S602. Preset error threshold: When the error output by the error calculation model exceeds the error threshold, the prediction model is retrained and the physical data in the digital twin model is updated.

[0021] A heat pump energy optimization operation system for use in livestock farms includes a data acquisition module, a digital twin model module, a prediction module, a variable optimization module, a control module, and a model optimization module. The data acquisition module is used to build a heat pump energy application system in the farm, and to collect working data using different sensors when the heat pump energy application system is working. The digital twin model module is used to collect physical data of the heat pump energy application system and construct a digital twin model of the heat pump energy application system based on the physical data and operating data. The prediction module is used to build a prediction model using machine learning algorithms, train the prediction model using historical working data of heat pump energy application systems, and predict future heat load and waste heat through the prediction model. The variable optimization module is used to calculate the comprehensive energy efficiency ratio of the heat pump energy application system using working data, set constraints, and use model predictive control algorithm to solve the optimal operating variables of the heat pump energy application system. The control module is used to generate control commands using optimal operating variables, and to adjust the operating status of the heat pump energy application system using the control commands. The model optimization module is used to collect the predicted and actual values ​​in the prediction model during the real-time operation of the heat pump energy application system, and to calculate the prediction error to optimize the model.

[0022] The digital twin model module includes a heat pump heating model unit, a heat exchanger model unit, and a pipeline heat loss model unit; The heat pump heating model unit is used to extract the heat pump performance coefficient from the heat pump and to construct a heat pump heating model using the heat pump performance coefficient and heat pump power consumption. The heat exchanger model unit is used to calculate the heat exchanger efficiency by combining the number of heat transfer units with the heat capacity flow rate ratio, and to calculate the actual heat exchange through the heat exchanger efficiency and the maximum heat exchange. The pipeline heat loss model unit is used to collect the surface area, heat transfer coefficient and water temperature of each connecting pipeline in the heat pump energy application system, and to calculate the pipeline heat loss.

[0023] The prediction module includes a heat load prediction unit and a waste heat prediction unit; The heat load prediction unit is used to train the heat load prediction model by using historical production plans, fermenter temperatures, ambient temperatures, and fermenter heat loads as training data. The real-time production plans, fermenter temperatures, and ambient temperatures are input into the heat load prediction model to predict the fermenter heat load for the next hour. The waste heat prediction unit is used to train the waste heat prediction model by using biogas slurry flow rate, biogas slurry temperature, actual heat exchanger heat exchanger heat exchanger heat exchanger and weather data as training data. The real-time biogas slurry temperature, biogas slurry flow rate and weather data are input into the waste heat prediction model to predict the actual heat exchanger ...

[0024] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention achieves efficient utilization of the entire energy chain through the synergy of "waste heat recovery + predictive optimization + on-demand control": compared with traditional systems, the overall energy efficiency ratio (SCOP) is significantly improved and the annual energy consumption cost is greatly reduced.

[0025] 2. This invention integrates four core technologies: Internet of Things, digital twin, machine learning, and model predictive control, to achieve "full-process automation and intelligence": without the need for manual data recording or equipment adjustment based on experience, the system can autonomously complete "data acquisition - model calculation - prediction optimization - control execution - model update", reducing labor costs and avoiding the "subjectivity and error" of manual operation. Attached Figure Description

[0026] Fig. 1 This is a schematic diagram of the heat pump energy application system structure of a heat pump energy optimization operation system applied to a breeding farm according to the present invention; Fig. 2 This is a module distribution diagram of a heat pump energy optimization operation system applied to a livestock farm according to the present invention. Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] Example: Figs. 1-2 As shown, the present invention provides a technical solution. A method for optimizing the energy operation of a heat pump in a livestock farm, the method comprising the following steps: S100. Construct a heat pump energy application system in the breeding farm, and collect working data using different sensors when the heat pump energy application system is working. The specific steps for collecting heat source, load, environmental, and device operation data using different sensors as working data in a heat pump energy application system are as follows: S101. Construct a heat pump energy application system in a breeding farm. The heat pump energy application system includes a biogas slurry tank, a heat exchanger, a heat pump, a boiler, and a fermentation tank. The heat pump includes an evaporator and a condenser. The biogas slurry tank transfers waste heat to the heat exchanger. The heat exchanger uses the waste heat to generate hot water, which enters the heat pump. The heat pump reheats the hot water and inputs it through the fermentation tank. While the heat pump is reheating the hot water, the boiler provides supplemental heat to the heat pump. All devices are connected through pipes. S102. When the heat pump energy application system is working, different sensors are used to collect working data, including biogas slurry temperature, biogas slurry flow rate, fermentation tank temperature, ambient temperature, heat pump power consumption, and boiler heat supplementation; the sensors transmit the collected data to the central processing unit through the Internet of Things.

[0029] It covers comprehensive data on heat sources, loads, environment, and equipment operation, providing a precise data source for subsequent model building and optimization. Data transmission via the Internet of Things ensures real-time data transmission, supporting subsequent real-time control decisions.

[0030] S200: Collect physical data of the heat pump energy application system, and construct a digital twin model of the heat pump energy application system based on the physical data and operating data; construct sub-models in the digital twin model, including a heat pump heating model, a heat exchanger model, and a pipeline heat loss model; The specific steps for constructing a digital twin model of a heat pump energy application system are as follows: S201. Collect physical data of the heat pump energy application system, use the physical data to obtain a digital twin model of the heat pump energy application system, and construct sub-models in the digital twin model. The sub-models include a heat pump heating model, a heat exchanger model, and a pipeline heat loss model. By breaking down the heat pump into three sub-models—heat pump heating, heat exchanger, and pipeline heat loss—the physical laws governing device operation can be accurately replicated, enabling visualization of the operating status.

[0031] The method for constructing a heat pump heating model is as follows: Extract the coefficient of performance (COP) from the heat pump, and then construct a heat pump heating model using the COP and heat pump power consumption. The formula is: ; In the formula, Q heat The heat pump's heating capacity is represented by W, and COP represents the coefficient of performance of the heat pump. hp Indicates the heat pump power consumption; Traditional systems can only roughly determine the operating status of a heat pump through "output temperature," and cannot quantify the "matching relationship between energy consumption and heating capacity." This model can accurately identify situations where "COP decreases but W..." hp The abnormal operating condition of "rising" can provide early warning of equipment failure, avoid energy waste caused by inefficient equipment operation, and provide "quantitative basis" for subsequent optimization of heat pump operating variables.

[0032] The model parameters can be updated in real time to adapt to the "performance degradation" of heat pumps during long-term operation. Compared with traditional fixed parameter models, it can more realistically reflect the actual operating status of the equipment and improve simulation accuracy.

[0033] S202. The method for constructing the heat exchanger model is as follows: Extract the mass flow rate and specific heat capacity at constant pressure of the heat exchanger. Multiply the mass flow rate by the specific heat capacity at constant pressure and then by the temperature difference to obtain the heat transfer capacity. Extract the maximum and minimum heat transfer capacities. Calculate the heat exchanger's heat capacity flow rate ratio r by dividing the minimum heat transfer capacity ratio by the maximum heat transfer capacity ratio. Calculate the heat exchanger efficiency using the number of heat transfer units method combined with the heat capacity flow rate ratio. Calculate the actual heat transfer capacity using the heat exchanger efficiency and the maximum heat transfer capacity. The formula is: ; In the formula, Q sh Indicates the actual heat exchange. Q represents the heat exchanger efficiency. max Indicates the maximum heat exchange capacity; Traditional systems monitor heat exchangers only based on the "inlet and outlet temperature difference," which cannot accurately calculate the actual heat exchange, and is prone to problems such as "insufficient waste heat recovery" or "excessive energy consumption." This model calculates the maximum heat exchange using "mass flow rate × specific heat capacity at constant pressure × temperature difference," and then combines this with heat capacity flow rate ratio optimization calculation to accurately pinpoint the "energy efficiency bottleneck" of the heat exchanger. For example, when the r value is too large, it indicates that the heat capacity flow rates of the cold and hot sides are mismatched. The flow rate can be optimized by adjusting the water pump frequency to improve the heat exchanger efficiency, making the actual heat exchange close to the maximum heat exchange, maximizing the waste heat recovery efficiency of biogas slurry, and reducing the dual problems of "waste heat" and "ineffective energy consumption."

[0034] S203. The method for constructing the pipeline heat loss model is as follows: In the heat pump energy application system, collect the surface area, heat transfer coefficient, and water temperature of each connecting pipe, and calculate the pipeline heat loss using the following formula: ; In the formula, Q loss U represents the heat loss of the pipe, A represents the heat transfer coefficient, and T represents the surface area of ​​the pipe. water Indicates water temperature, T amb It represents the ambient temperature; the real-time calculated values ​​of the three sub-models in the digital twin model are output as the operating data of the heat pump energy application system.

[0035] By collecting the heat transfer coefficient, surface area, water temperature, and ambient temperature of each pipe, the heat loss of the pipe is quantified segment by segment: Traditional systems often ignore the heat loss of the pipe and only calculate it based on "heat output of heat pump = heat obtained by fermentation tank", which leads to blindly increasing the heat pump power consumption when the actual heating is insufficient. This model can accurately calculate the heat loss of each pipe segment, providing a basis for subsequent "pipe insulation optimization" and "heating compensation", and further reducing energy waste. By integrating the real-time calculated values ​​of heat pump heating, heat exchanger heat exchange, and pipeline heat loss into "virtual operating data," and comparing them with the actual operating data of the physical device, "abnormal points" in the physical system can be intuitively identified, enabling "early fault identification and precise problem location," and reducing downtime losses in traditional systems where "faults can only be investigated after the fact."

[0036] By calculating key parameters using quantitative formulas, a reliable virtual simulation foundation is provided for subsequent prediction and optimization.

[0037] S300: Utilize machine learning algorithms to build a prediction model, train the prediction model using historical heat pump energy application system operating data, and predict future heat load and waste heat through the prediction model; The specific steps for predicting future heat load and waste heat using a predictive model are as follows: S301. Construct a heat load prediction model and a waste heat prediction model using machine learning algorithms. For the heat load prediction model, extract the heat pump heating capacity, boiler supplementary heating capacity, and pipeline heat loss from the historical digital twin model. Use the heat pump heating capacity plus the boiler supplementary heating capacity and then subtract the pipeline heat loss to obtain the fermenter heat load. Extract historical production plans, fermenter temperatures, and ambient temperatures. Use the historical production plans, fermenter temperatures, ambient temperatures, and fermenter heat loads as training data to train the heat load prediction model. Set the prediction time to 1 hour. Input the real-time production plans, fermenter temperatures, and ambient temperatures into the heat load prediction model to predict the fermenter heat load for the next 1 hour. S302. For the waste heat prediction model, extract historical data on biogas slurry flow rate, biogas slurry temperature, actual heat exchange of the heat exchanger, and weather data. Use the biogas slurry flow rate, biogas slurry temperature, actual heat exchange of the heat exchanger, and weather data as training data to train the waste heat prediction model. Set the prediction time to 1 hour, input the real-time biogas slurry temperature, biogas slurry flow rate, and weather data into the waste heat prediction model to predict the actual heat exchange of the heat exchanger for the next 1 hour, and use the actual heat exchange of the heat exchanger as the waste heat of the biogas slurry.

[0038] Staff members input production plans and weather data quantitatively based on experience, converting them into fixed values. S400: Calculate the overall energy efficiency ratio of the heat pump energy application system using working data, set constraints, and use the model predictive control algorithm to solve for the optimal operating variables of the heat pump energy application system. The specific steps for solving the optimal operating variables of a heat pump energy application system using the model predictive control algorithm are as follows: S401. Input the predicted heat load and predicted waste heat for the next hour into the digital twin model to extract the heat pump heating capacity, heat pump power consumption, and total system power consumption of the heat pump energy application system for the next hour. Calculate the comprehensive energy efficiency ratio of the heat pump energy application system using the following formula: ; In the formula, SCOP represents the overall energy efficiency ratio of the heat pump energy application system, W. hp This indicates the heat pump's power consumption, in W. pump This indicates the total system power consumption of a heat pump energy application system; Compared to the traditional single indicator of "only calculating the COP of heat pumps", SCOP covers the total power consumption of all devices such as heat pumps and water pumps, and can better reflect the overall energy utilization efficiency of the system. For example, under a certain operating condition, the COP of the heat pump is high, but the water pump has a large total power consumption due to excessive frequency. Traditional indicators will misjudge "the system has good energy efficiency", while SCOP can accurately identify the problem of "local high efficiency but overall low efficiency", providing a "global perspective" for optimization.

[0039] S402. In the digital twin model, under the standard of predicted heat load and predicted waste heat for the next hour, the heat pump condensing temperature, evaporating temperature, cold-side water pump frequency and hot-side water pump frequency in the heat pump energy application system are adjusted. After each adjustment, the corresponding comprehensive energy efficiency ratio is calculated, and the heat pump condensing temperature, evaporating temperature, cold-side water pump frequency and hot-side water pump frequency corresponding to the maximum comprehensive energy efficiency ratio are selected as the optimal operating variables. S403. When debugging the heat pump condensing temperature, evaporation temperature, cold-side water pump frequency and hot-side water pump frequency in the heat pump energy application system, set constraints. Specifically, in the digital twin model, based on professional experience, the staff sets the operating limits of the heat pump energy application system according to the physical laws of the three sub-models. The operating limits are the maximum values ​​that the fermenter temperature, operating pressure and system power can withstand when the heat pump energy application system is running.

[0040] With the goal of maximizing SCOP in the next hour, the four core operating variables of heat pump condensing temperature, evaporating temperature, cold-side water pump frequency, and hot-side water pump frequency were adjusted. Traditional control often adopts "trial and error with a single variable", which can easily lead to "local optimization but overall inefficiency". However, the model predictive control algorithm can consider the interaction of the four variables at the same time. Through multiple rounds of simulation and debugging, the optimal combination of "heating capacity meets demand and total power consumption is minimized" is selected.

[0041] The algorithm is highly efficient, completing a round of optimization calculations within 10-20 seconds. It is adapted to the control requirements of a "1-hour prediction cycle", ensuring that the optimal operating variables can be applied to actual operation in a timely manner and avoiding "optimization failure due to calculation lag".

[0042] S500: Generates control commands using optimal operating variables, and uses control commands to adjust the operating status of the heat pump energy application system; The specific steps for adjusting the operating status of a heat pump energy application system using control commands are as follows: S501: The real-time heat pump condensing temperature, evaporating temperature, cold-side water pump frequency, and hot-side water pump frequency in the heat pump energy application system are collected and compared with the optimal operating variables. The difference between the four operating variables is calculated, and the difference is converted into a control command and sent to the corresponding device for control. The control command is issued through the PLC. After the heat pump energy application system is adjusted by the control command of the optimal operating variables, it returns to the working data of the heat pump energy application system collected in S100, and then executes the subsequent steps to form a closed loop. S502. Extract the heat pump's heating capacity and the fermenter's heat load. When the heat pump's heating capacity is less than the fermenter's heat load, start the boiler to supplement heat. S600 collects predicted and actual values ​​from the prediction model during real-time operation of the heat pump energy application system, and calculates the prediction error to optimize the model.

[0043] The specific steps for calculating the prediction error optimization model are as follows: S601. In the operation of a heat pump energy application system, a model performance evaluation is performed weekly. Operating data and actual operating data are extracted from the digital twin model, and predicted values ​​and actual values ​​are extracted from the prediction model. These are then input into the error calculation model to calculate the errors of the digital twin model and the prediction model. The preset error calculation model is as follows: ; In the formula, Error represents the model error, measured represents the data value in the model, corresponding to the working data in the digital twin model and the predicted value in the prediction model, and predicted represents the actual measured value, corresponding to the actual working data and the actual value. S602. Preset error threshold: When the error output by the error calculation model exceeds the error threshold, the prediction model is retrained and the physical data in the digital twin model is updated.

[0044] A heat pump energy optimization operation system for use in livestock farms includes a data acquisition module, a digital twin model module, a prediction module, a variable optimization module, a control module, and a model optimization module. The data acquisition module is used to build a heat pump energy application system in the farm, and to collect working data using different sensors when the heat pump energy application system is working. The digital twin model module is used to collect physical data of the heat pump energy application system and construct a digital twin model of the heat pump energy application system based on the physical data and operating data. The prediction module is used to build a prediction model using machine learning algorithms, train the prediction model using historical working data of heat pump energy application systems, and predict future heat load and waste heat through the prediction model. The variable optimization module is used to calculate the comprehensive energy efficiency ratio of the heat pump energy application system using working data, set constraints, and use model predictive control algorithm to solve the optimal operating variables of the heat pump energy application system. The control module is used to generate control commands using optimal operating variables, and to adjust the operating status of the heat pump energy application system using the control commands. The model optimization module is used to collect the predicted and actual values ​​in the prediction model during the real-time operation of the heat pump energy application system, and to calculate the prediction error to optimize the model.

[0045] The digital twin model module includes a heat pump heating model unit, a heat exchanger model unit, and a pipeline heat loss model unit; The heat pump heating model unit is used to extract the heat pump performance coefficient from the heat pump and to construct a heat pump heating model using the heat pump performance coefficient and heat pump power consumption. The heat exchanger model unit is used to calculate the heat exchanger efficiency by combining the number of heat transfer units with the heat capacity flow rate ratio, and to calculate the actual heat exchange through the heat exchanger efficiency and the maximum heat exchange. The pipeline heat loss model unit is used to collect the surface area, heat transfer coefficient and water temperature of each connecting pipeline in the heat pump energy application system, and to calculate the pipeline heat loss.

[0046] The prediction module includes a heat load prediction unit and a waste heat prediction unit; The heat load prediction unit is used to train the heat load prediction model by using historical production plans, fermenter temperatures, ambient temperatures, and fermenter heat loads as training data. The real-time production plans, fermenter temperatures, and ambient temperatures are input into the heat load prediction model to predict the fermenter heat load for the next hour. The waste heat prediction unit is used to train the waste heat prediction model by using biogas slurry flow rate, biogas slurry temperature, actual heat exchanger heat exchanger heat exchanger heat exchanger and weather data as training data. The real-time biogas slurry temperature, biogas slurry flow rate and weather data are input into the waste heat prediction model to predict the actual heat exchanger ... Example

[0047] A large-scale pig farm with 10,000 pigs is located in North China, mainly raising fattening pigs and breeding sows. It needs to meet two core heating needs daily: first, the fermentation treatment of pig manure in fermentation tanks (maintaining a tank temperature of 35-40℃ to ensure harmless treatment of manure and efficient biogas production); and second, winter heating for pigsties (maintaining a pigsty temperature of 18-22℃ to ensure a suitable environment for pig growth). The heat pump energy application system described in this invention is installed in the farm. A temperature-flow integrated sensor (measurement range 0-50℃, 0-20m³ / h, accuracy ±0.5℃, ±0.1m³ / h) is installed at the outlet of the biogas slurry buffer tank. A multi-point temperature sensor (3 measuring points to avoid local temperature deviation) is installed inside the fermentation tank. Pressure-temperature sensors are installed at the inlet and outlet of the heat pump. A calorimeter (measuring supplemental heat) is installed at the boiler outlet. An ambient temperature sensor is installed in the pigsty and outdoors. A total of 12 sensors are deployed, covering the entire chain of "waste heat-heat exchange-heat generation-heat use-environment".

[0048] Data transmission system setup: LoRa wireless transmission modules (suitable for open environments in farms, with a transmission distance of up to 1km) are used to upload sensor data to the central processor in real time (an industrial-grade PLC is selected, supporting data storage and edge computing). The data update frequency is 1 minute / time. At the same time, a visual monitoring platform (web terminal + mobile terminal) is built, allowing managers to view the operating parameters and data curves of each device in real time.

[0049] Collect core physical data, including the heat pump's rated power (15.6kW), COP curve, heat exchanger's heat transfer coefficient (1200W / (㎡・℃)), isobaric specific heat capacity (water's isobaric specific heat capacity is 4.186kJ / (kg・℃)), pipe length (total length approximately 300m), pipe diameter (DN80 pipe 150m, DN100 pipe 150m), surface area (calculated total surface area approximately 141.3㎡), heat transfer coefficient (after insulation, U=0.2W / (㎡・℃)), fermenter volume (100m³), heat dissipation coefficient (5W / (㎡・℃)), etc., to ensure that the data is consistent with the actual equipment parameters.

[0050] Three sub-models were constructed based on the collected data to obtain a heat pump heating model, a heat exchanger model, and a pipeline heat loss model. These three sub-models were then connected to a digital twin platform to output key parameters such as "heat pump heating capacity, heat exchanger heat exchange capacity, and pipeline heat loss" in real time. These parameters were compared with the physical equipment operating data in real time to form a "virtual-real mapping". Managers could view the equipment operating status through the platform. If they found that "the virtual heat exchange capacity is much lower than the actual value", they could promptly investigate the problem of heat exchanger scaling.

[0051] A predictive model was constructed using machine learning algorithms. Taking the heat load prediction model as an example, an LSTM neural network algorithm was used. The input features were "production plan (quantized as 0-1 variables, where 1 represents the fermenter operating at full load), fermenter temperature, and ambient temperature," and the output was "fermenter heat load for the next hour." After training, the model's prediction error rate on the test set was 6.2%. For example, when the real-time production plan was 1, the fermenter temperature was 37℃, and the ambient temperature was 5℃, the predicted heat load for the next hour was 42kW, while the actual value was 40.5kW, a deviation of 1.5kW.

[0052] The predictive model is deployed to the central processor, generating an hourly "Predicted Heat Load and Waste Heat Report for the Next Hour" and pushing it to the visualization platform. For example, at 7 a.m. on a winter morning, if the forecast predicts that the ambient temperature will drop to -8°C, the heat load will rise to 145kW, and the waste heat will drop to 125kW in the next hour, the system will issue an early warning that "the waste heat is insufficient, and it is necessary to prepare for heat pump efficiency improvement or boiler supplemental heating."

[0053] The overall energy efficiency ratio (SCOP) was then calculated using a Model Predictive Control (MPC) algorithm. With the objective of maximizing the SCOP over the next hour, the algorithm solved for four key operating variables: heat pump condensing temperature, evaporating temperature, cold-side pump frequency, and hot-side pump frequency. For example, when the predicted heat load for the next hour was 145kW and the waste heat was 125kW, the algorithm, through multiple rounds of simulation and debugging, selected the optimal operating variables: condensing temperature 42℃, evaporating temperature -5℃, cold-side pump frequency 40Hz, and hot-side pump frequency 38Hz. At this point, the calculated SCOP was 1.35, a 32.3% improvement compared to the random adjustment variable (SCOP = 1.02).

[0054] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A method for optimizing the energy operation of a heat pump applied in a livestock farm, characterized in that: The method includes the following steps: S100. Construct a heat pump energy application system in the breeding farm, and collect working data using different sensors when the heat pump energy application system is working. S200: Collect physical data of the heat pump energy application system, and construct a digital twin model of the heat pump energy application system based on the physical data and operating data; construct sub-models in the digital twin model, including a heat pump heating model, a heat exchanger model, and a pipeline heat loss model; S300: Utilize machine learning algorithms to build a prediction model, train the prediction model using historical heat pump energy application system operating data, and predict future heat load and waste heat through the prediction model; S400: Calculate the overall energy efficiency ratio of the heat pump energy application system using working data, set constraints, and use the model predictive control algorithm to solve for the optimal operating variables of the heat pump energy application system. S500: Generates control commands using optimal operating variables, and uses control commands to adjust the operating status of the heat pump energy application system; S600 collects predicted and actual values ​​from the prediction model during real-time operation of the heat pump energy application system, and calculates the prediction error to optimize the model.

2. The method for optimizing the operation of a heat pump in a livestock farm according to claim 1, characterized in that: The specific steps in S100 for collecting heat source, load, environmental, and device operation data from different sensors as working data during the operation of the heat pump energy application system are as follows: S101. Construct a heat pump energy application system in a breeding farm. The heat pump energy application system includes a biogas slurry tank, a heat exchanger, a heat pump, a boiler, and a fermentation tank. The heat pump includes an evaporator and a condenser. The biogas slurry tank provides waste heat. The heat exchanger transfers the waste heat from the biogas slurry tank to the circulating water on the evaporator side of the heat pump. The heat pump uses the heat from the circulating water on the evaporator side to generate high-temperature hot water on the condenser side through electrical energy. The high-temperature hot water is input into the fermentation tank. When the heat pump generates high-temperature hot water, the boiler provides supplemental heat to the heat pump. All devices are connected through pipes. S102. When the heat pump energy application system is working, different sensors are used to collect working data, including biogas slurry temperature, biogas slurry flow rate, fermentation tank temperature, ambient temperature, heat pump power consumption, and boiler heat supplementation; the sensors transmit the collected data to the central processing unit through the Internet of Things.

3. The method for optimizing the operation of a heat pump in a livestock farm according to claim 2, characterized in that: The specific steps for constructing a digital twin model of the heat pump energy application system in S200 are as follows: S201. Collect physical data of the heat pump energy application system, use the physical data to obtain a digital twin model of the heat pump energy application system, and construct sub-models in the digital twin model. The sub-models include a heat pump heating model, a heat exchanger model, and a pipeline heat loss model. The method for constructing a heat pump heating model is as follows: extract the heat pump performance coefficient from the heat pump, and construct a heat pump heating model using the heat pump performance coefficient and heat pump power consumption. S202. The method for constructing the heat exchanger model is as follows: extract the mass flow rate and specific heat capacity at constant pressure of the heat exchanger, multiply the mass flow rate by the specific heat capacity at constant pressure and then by the temperature difference to obtain the heat transfer capacity, and extract the maximum and minimum heat transfer capacity; use the ratio of the minimum heat transfer capacity to the maximum heat transfer capacity to obtain the heat exchanger heat capacity flow rate ratio r, use the number of heat transfer units method combined with the heat capacity flow rate ratio to calculate the heat exchanger efficiency, and use the heat exchanger efficiency and the maximum heat transfer capacity to calculate the actual heat transfer capacity. S203. The method for constructing the pipeline heat loss model is as follows: collect the surface area, heat transfer coefficient and water temperature of each connecting pipeline in the heat pump energy application system, and calculate the pipeline heat loss. The real-time calculated values ​​of the three sub-models in the digital twin model are output as the operating data of the heat pump energy application system.

4. The method for optimizing the operation of a heat pump in a livestock farm according to claim 3, characterized in that: The specific steps in S300 for predicting future heat load and waste heat using a prediction model are as follows: S301. Construct a heat load prediction model and a waste heat prediction model using machine learning algorithms. Extract the heat pump heating capacity, boiler supplementary heating capacity, and pipeline heat loss from the historical digital twin model for the heat load prediction model. Use the heat pump heating capacity plus the boiler supplementary heating capacity and subtract the pipeline heat loss to obtain the fermenter heat load. Extract historical production plans, fermenter temperatures, and ambient temperatures. Use historical production plans, fermenter temperatures, ambient temperatures, and fermenter heat loads as training data to train the heat load prediction model. Set the prediction time to 1 hour. Input the real-time production plan, fermenter temperature, and ambient temperature into the heat load prediction model to predict the fermenter heat load for the next 1 hour. S302. For the waste heat prediction model, extract historical biogas slurry flow rate, biogas slurry temperature, actual heat exchanger heat transfer, and weather data. Use the biogas slurry flow rate, biogas slurry temperature, actual heat exchanger heat transfer, and weather data as training data to train the waste heat prediction model. Set the prediction time to 1 hour, input the real-time biogas slurry temperature, biogas slurry flow rate, and weather data into the waste heat prediction model to predict the actual heat exchanger heat transfer in the next 1 hour, and use the actual heat exchanger heat transfer as the waste heat of the biogas slurry.

5. The method for optimizing the operation of a heat pump in a livestock farm according to claim 4, characterized in that: The specific steps in S400 for solving the optimal operating variables of the heat pump energy application system using the model predictive control algorithm are as follows: S401. Input the predicted heat load and predicted waste heat for the next hour into the digital twin model to extract the heat pump heating capacity, heat pump power consumption and total system power consumption of the heat pump energy application system for the next hour, and calculate the comprehensive energy efficiency ratio of the heat pump energy application system. S402. In the digital twin model, under the standard of predicted heat load and predicted waste heat for the next hour, the heat pump condensing temperature, evaporating temperature, cold-side water pump frequency and hot-side water pump frequency in the heat pump energy application system are adjusted. After each adjustment, the corresponding comprehensive energy efficiency ratio is calculated, and the heat pump condensing temperature, evaporating temperature, cold-side water pump frequency and hot-side water pump frequency corresponding to the maximum comprehensive energy efficiency ratio are selected as the optimal operating variables. S403. When debugging the heat pump condensing temperature, evaporation temperature, cold-side water pump frequency and hot-side water pump frequency in the heat pump energy application system, set constraints. Specifically, in the digital twin model, based on professional experience, the staff sets the operating limits of the heat pump energy application system according to the physical laws of the three sub-models. The operating limits are the maximum values ​​that the fermenter temperature, operating pressure and system power can withstand when the heat pump energy application system is running.

6. The method for optimizing the operation of a heat pump in a livestock farm according to claim 5, characterized in that: The specific steps for adjusting the operating status of the heat pump energy application system using control commands in S500 are as follows: S501: The real-time heat pump condensing temperature, evaporating temperature, cold-side water pump frequency, and hot-side water pump frequency in the heat pump energy application system are collected and compared with the optimal operating variables. The difference between the four operating variables is calculated, and the difference is converted into a control command and sent to the corresponding device for control. The control command is issued through the PLC. After the heat pump energy application system is adjusted by the control command of the optimal operating variables, it returns to the working data of the heat pump energy application system collected in S100, and then executes the subsequent steps to form a closed loop. S502. Extract the heat output of the heat pump and the heat load of the fermenter. When the heat output of the heat pump is less than the heat load of the fermenter, start the boiler to supplement the heat.

7. The method for optimizing the operation of a heat pump in a livestock farm according to claim 6, characterized in that: The specific steps for calculating the prediction error optimization model in S600 are as follows: S601. In the operation of a heat pump energy application system, a model performance evaluation is performed weekly. Operating data and actual operating data are extracted from the digital twin model, and predicted values ​​and actual values ​​are extracted from the prediction model. These are then input into the error calculation model to calculate the errors of the digital twin model and the prediction model. The preset error calculation model is as follows: ； In the formula, Error represents the model error, measured represents the data value in the model, corresponding to the working data in the digital twin model and the predicted value in the prediction model, and predicted represents the actual measured value, corresponding to the actual working data and the actual value. S602. Preset error threshold: When the error output by the error calculation model exceeds the error threshold, the prediction model is retrained and the physical data in the digital twin model is updated.

8. A heat pump energy optimization operation system applied to a livestock farm, characterized in that: The heat pump energy optimization operation system includes a data acquisition module, a digital twin model module, a prediction module, a variable optimization module, a control module, and a model optimization module; The data acquisition module is used to build a heat pump energy application system in the farm, and to collect working data using different sensors when the heat pump energy application system is working. The digital twin model module is used to collect physical data of the heat pump energy application system and construct a digital twin model of the heat pump energy application system based on the physical data and operating data. The prediction module is used to build a prediction model using machine learning algorithms, train the prediction model using historical working data of heat pump energy application systems, and predict future heat load and waste heat through the prediction model. The variable optimization module is used to calculate the comprehensive energy efficiency ratio of the heat pump energy application system using working data, set constraints, and use model predictive control algorithm to solve the optimal operating variables of the heat pump energy application system. The control module is used to generate control commands using optimal operating variables, and to adjust the operating status of the heat pump energy application system using the control commands. The model optimization module is used to collect the predicted and actual values ​​in the prediction model during the real-time operation of the heat pump energy application system, and to calculate the prediction error to optimize the model.

9. A heat pump energy optimization operation system for a livestock farm according to claim 8, characterized in that: The digital twin model module includes a heat pump heating model unit, a heat exchanger model unit, and a pipeline heat loss model unit. The heat pump heating model unit is used to extract the heat pump performance coefficient from the heat pump and to construct a heat pump heating model using the heat pump performance coefficient and heat pump power consumption. The heat exchanger model unit is used to calculate the heat exchanger efficiency by combining the number of heat transfer units with the heat capacity flow rate ratio, and to calculate the actual heat exchange through the heat exchanger efficiency and the maximum heat exchange. The pipeline heat loss model unit is used to collect the surface area, heat transfer coefficient and water temperature of each connecting pipeline in the heat pump energy application system, and to calculate the pipeline heat loss.

10. A heat pump energy optimization operation system for a livestock farm according to claim 8, characterized in that: The prediction module includes a heat load prediction unit and a waste heat prediction unit; The heat load prediction unit is used to train the heat load prediction model by using historical production plans, fermenter temperatures, ambient temperatures, and fermenter heat loads as training data. The real-time production plans, fermenter temperatures, and ambient temperatures are input into the heat load prediction model to predict the fermenter heat load for the next hour. The waste heat prediction unit is used to train the waste heat prediction model by using biogas slurry flow rate, biogas slurry temperature, actual heat exchange of the heat exchanger, and weather data as training data. The real-time biogas slurry temperature, biogas slurry flow rate, and weather data are input into the waste heat prediction model to predict the actual heat exchange of the heat exchanger for the next hour, and the actual heat exchange of the heat exchanger is used as the waste heat of the biogas slurry.

Images