A liquid level and temperature control system for an industrial chiller

Abstract

This invention relates to the field of precision temperature control technology and discloses a liquid level and temperature control system for an industrial chiller, comprising: a liquid level adjustment loop, a temperature adjustment loop, and a collaborative control device with a unidirectional logic channel. The system acquires the temperature parameters of the makeup water medium and the circulating medium, retrieves the estimated mass flow rate by calling a preset flow image model, calculates the instantaneous enthalpy-flow change rate introduced by the makeup water action, converts it into a feedforward compensation signal superimposed on the temperature adjustment loop, and executes model parameter self-calibration logic based on thermal response residuals. This invention eliminates physical mixing hysteresis by establishing a logical hard link for the hydrothermal action, achieving zero-time-difference immunity to makeup water thermal disturbances; simultaneously, it uses temperature residuals to inversely correct the flow model parameters, improving the system's adaptability to water pressure fluctuations and component aging under flow meter-less conditions.

Description

Technical Field

[0001] This invention relates to a liquid level and temperature control system for an industrial chiller, belonging to the field of precision temperature control technology. Background Technology

[0002] Current industrial chiller control architectures generally employ independent closed-loop control modes for liquid level and temperature. The liquid level regulation loop drives the opening and closing of the water replenishment actuator based on the liquid level detection signal to maintain the working fluid capacity, while the temperature regulation loop adjusts the output power of the heat exchange actuator based on the circulating medium temperature detection signal to maintain thermal balance. This single-variable independent feedback control mode is simple in structure and meets basic cooling requirements under normal steady-state conditions. However, in high-power lasers or semiconductor manufacturing and other precision cooling scenarios sensitive to temperature fluctuations, the aforementioned independent control architecture faces constraints such as physical transport lag and siloed logical information. The water replenishment action requires the introduction of specific... For fresh media with temperature differences, the temperature sensor experiences a physical delay in sensing thermal disturbances due to pipeline transport distance and fluid mixing limitations. The temperature regulation loop only intervenes after the temperature of the probe changes due to fluid mixing, causing the energy compensation action to lag behind the actual heat load change and leading to water temperature overshoot oscillation. Existing improvement solutions usually involve increasing the volume of the buffer tank to dilute the thermal shock or adding a flow meter to build a hardware feedforward loop. The former increases the size of the equipment and manufacturing cost, while the latter cannot solve the problem of phase synchronization between the control signal and the arrival time of the heat cluster under variable flow rate conditions, and lacks dynamic constraints on the physical capability boundaries of the actuator.

[0003] While addressing the lag in physical transport, existing technologies have failed to establish a deterministic correlation between liquid level action and temperature response at the control logic level. They still rely on passive and easily disturbed parameter models. For example, Chinese invention patent CN115201657A discloses a method and device for controlling a chiller unit, a chiller unit, and a storage medium. This solution provides control and regulation based on temperature detection, but the core logic still relies on feedback or feedforward compensation of the chiller unit's operating load. It does not construct a high-precision enthalpy-flow decoupling model with a deterministic correlation to mass flow to address the sudden thermal disturbances introduced by the water replenishment action. Crucially, this solution relies on pre-calibrated parameters for control and lacks an adaptive correction mechanism for fluctuations in the water replenishment network pressure, wear of actuators, or drift in the water tank's liquid level characteristics. Once the on-site operating conditions change, the accuracy of flow estimation drops sharply, and the compensation amount will inevitably be under-compensated or over-compensated, ultimately causing continuous oscillations or deviations in system temperature, making it difficult to achieve endogenous adaptability to fluid parameter drift.

[0004] Therefore, the technical problem to be solved by this invention is how to establish a deterministic correlation between liquid level action and temperature response without relying on external flow detection hardware, and achieve precise suppression of thermal disturbances in water replenishment while overcoming fluid transport lag and actuator physical constraints. Summary of the Invention

[0005] To address the problems mentioned in the background art, the technical solution of the present invention is as follows: A liquid level and temperature control system for an industrial chiller, comprising a liquid level regulating loop, a temperature regulating loop, and a collaborative control device, wherein the collaborative control device establishes a unidirectional logic channel connecting the liquid level regulating loop and the temperature regulating loop, and the collaborative control device executes the following control logic:

[0006] Within the synchronous clock cycle of the water replenishment opening command output by the liquid level regulation loop, the water replenishment medium temperature parameters and the circulating medium temperature parameters are obtained. The flow image model pre-installed in the cooperative control device is called to find the mass flow rate estimate corresponding to the water replenishment opening command. Based on the mass flow rate estimate, the instantaneous enthalpy change rate introduced by the water replenishment action is calculated. The instantaneous enthalpy change rate is converted into a feedforward compensation signal with the same dimensions as the temperature regulation loop and superimposed on the power control command of the temperature regulation loop.

[0007] The collaborative control device also executes model parameter self-calibration logic: within the preset evaluation time window after the water replenishment action corresponding to the water replenishment opening command ends, it calculates the integral of the deviation of the circulating medium temperature parameter relative to the temperature setpoint.

[0008] Based on the polarity of the deviation integral and the polarity of the difference between the temperature parameters of the makeup water medium and the temperature parameters of the circulating medium, it is determined whether the feedforward compensation of the previous makeup water cycle is under-compensated or over-compensated.

[0009] Based on the judgment results, the flow gain coefficient in the flow image model is modified step by step, and the modified flow gain coefficient is used to update the flow image model to participate in the mass flow estimation of the next water replenishment cycle, so as to offset the model error introduced by the pressure fluctuation of the water replenishment pipeline or the drift of the actuator characteristics.

[0010] Preferably, the collaborative control device further includes a timing shaping module, which operates a first-order inertial delay function. The collaborative control device processes the feedforward compensation signal using the first-order inertial delay function, and then superimposes the processed feedforward compensation signal onto the power control command, so that the power regulation response curve of the heat exchange actuator controlled by the temperature regulation loop is synchronized with the thermal diffusion response curve of the makeup water medium in the chiller tank in the time domain.

[0011] Preferably, the collaborative control device also establishes a communication connection with the circulating pump drive unit to obtain pump operating frequency parameters characterizing the flow velocity of the circulating medium; the timing shaping module operates dynamic delay mapping logic, which calculates the time constant of the first-order inertial delay function in real time according to the following formula: ,in, The time constant of the first-order inertial delay function. The preset pipeline transport characteristic coefficient, The pump operating frequency parameter is used; the timing shaping module dynamically assigns the time constant to the calculation result, so that the effective time of the feedforward compensation signal automatically drifts with the change of the circulating medium flow rate.

[0012] Preferably, the coordinated control device also executes disturbance shaping logic based on thermal margin: before responding to the water replenishment opening command, it obtains the current remaining power regulation margin of the temperature regulation loop; based on the temperature difference between the water replenishment medium temperature parameter and the circulating medium temperature parameter, it reversely converts the remaining power regulation margin into the maximum water replenishment flow threshold allowed to be introduced into the system at the current moment; it compares the planned flow rate mapped based on the water replenishment opening command with the maximum water replenishment flow threshold; if the planned flow rate exceeds the maximum water replenishment flow threshold, it performs amplitude limiting processing on the water replenishment opening command to ensure that the instantaneous enthalpy change rate introduced by the water replenishment action is always within the physical regulation capability range of the temperature regulation loop.

[0013] Preferably, the collaborative control device also performs trend prediction and buffering logic: monitoring the rate of decrease of the liquid level detection value in the liquid level regulation loop and estimating the remaining time to trigger the water replenishment action; when the remaining time is less than a preset threshold, comparing the temperature parameters of the water replenishment medium and the temperature parameters of the circulating medium to determine the polarity of the expected thermal shock; before the water replenishment action is triggered, sending a pre-bias signal to the temperature regulation loop according to the polarity, driving the temperature regulation loop to pre-adjust the temperature parameters of the circulating medium in the opposite direction to the expected thermal shock, and building a thermal potential energy buffer against thermal shock in the liquid storage tank.

[0014] Preferably, the calculation logic for the instantaneous enthalpy change rate is as follows: calculate the temperature difference between the temperature parameters of the makeup water medium and the temperature parameters of the circulating medium, calculate the product of the temperature difference, the estimated mass flow rate, and the preset specific heat capacity of the medium, and define the product as the instantaneous enthalpy change rate; the conversion logic for the feedforward compensation signal is as follows: divide the instantaneous enthalpy change rate by the energy efficiency ratio coefficient of the compressor in the temperature regulation loop to obtain the equivalent compressor power increment value as the feedforward compensation signal.

[0015] Preferably, the logic for determining whether the feedforward compensation in the previous water replenishment cycle is in an under-compensated or over-compensated state specifically includes: if the temperature parameter of the water replenishment medium is lower than that of the circulating medium and the deviation integral is positive, it is determined to be in an under-compensated state, indicating that the actual water replenishment flow rate is greater than the estimated mass flow rate; if the temperature parameter of the water replenishment medium is lower than that of the circulating medium and the deviation integral is negative, it is determined to be in an over-compensated state, indicating that the actual water replenishment flow rate is less than the estimated mass flow rate.

[0016] Preferably, the collaborative control device obtains the temperature parameters of the replenishing medium by reading data from a temperature sensor installed at the inlet of the chiller's replenishing water pipeline; the collaborative control device obtains the temperature parameters of the circulating medium by reading data from a temperature sensor installed at the outlet of the chiller; the liquid level regulation circuit includes a liquid level sensor and a replenishing solenoid valve, and the replenishing water opening command is used to control the duty cycle or opening percentage of the opening time of the replenishing solenoid valve.

[0017] Preferably, the logic of the step-by-step correction of the flow gain coefficient is as follows: when it is determined to be undercompensated, a preset step value is added to the current flow gain coefficient; when it is determined to be overcompensated, a preset step value is reduced to the current flow gain coefficient; the initial value of the flow gain coefficient is preset to 1.0, and upper and lower limit boundary values ​​that allow correction are set to prevent coefficient divergence caused by temperature sensor failure.

[0018] Preferably, the collaborative control device is integrated into a programmable logic controller or an embedded microcontroller, and the unidirectional logic channel is the data variable mapping relationship in the shared memory area inside the controller; the power control command of the temperature regulation loop is used to drive the variable frequency compressor or electronic expansion valve to regulate the cooling capacity output of the refrigeration system.

[0019] Compared with the prior art, the beneficial effects of the present invention are:

[0020] 1. In the liquid level and temperature control of the chiller, this system establishes a unidirectional logical mapping channel between the liquid level regulation loop and the temperature regulation loop. It directly converts the liquid level water replenishment actuator opening control command into an enthalpy flow rate signal characterizing the change in heat load, which is then superimposed on the temperature regulation loop as a feedforward quantity. The system itself issues control actions as a deterministic disturbance prediction source, and completes the energy supply and demand pre-balance calculation before the physical medium mixes and the temperature sensor responds. Based on the information flow decoupling strategy, it avoids the response lag caused by the fluid mixing time and sensor thermal inertia of traditional feedback control. It does not require additional hardware detection costs such as flow meters, eliminates the impact of water replenishment actions on water temperature, and ensures continuous temperature control.

[0021] 2. This system is configured with timing shaping logic that matches the mixing characteristics of the chiller's water tank. By collecting the operating frequency parameters of the circulating pump in real time to characterize the medium flow rate, the system dynamically adjusts the inertial delay time constant of the feedforward signal. The logic ensures that the power adjustment action of the heat exchange actuator is precisely synchronized with the moment when the water replenishment medium flows through the heat exchange area on the time axis. The timing of the control signal activation is dynamically locked with the physical medium transport speed, solving the problem of control timing misalignment caused by flow rate changes under variable frequency operation conditions, and avoiding artificial temperature oscillations caused by the compensation signal being ahead or behind.

[0022] 3. This system introduces thermal margin-based disturbance shaping logic, changing the traditional one-way disturbance response mode and establishing a reverse constraint mechanism for liquid level action based on temperature regulation capability. Before executing the water replenishment action, the system calculates the maximum mass flow rate threshold that the system can withstand at this moment based on the current remaining power margin of the temperature regulation unit and the water replenishment temperature difference. Based on this, the water replenishment command is limited to ensure that the thermal disturbance introduced into the system is always within the coverage of the physical capability of the actuator. This eliminates the risk of temperature runaway caused by the compressor or heater entering a saturated state from the bottom layer of the control logic, and improves the system's operational stability under extreme high temperature or full load conditions. Attached Figure Description

[0023] Figure 1 This is a flowchart of the collaborative control logic of enthalpy feedforward and model self-calibration in this invention;

[0024] Figure 2 This is a comparison curve of the temperature response between the collaborative control of the present invention under water supply pressure fluctuation and the traditional PID;

[0025] Figure 3 This is a schematic diagram of the hardware topology and signal interaction principle of the liquid level and temperature coordinated control system of the present invention. Detailed Implementation

[0026] This specific embodiment provides a liquid level and temperature control system for an industrial chiller. The following content is only used to explain and illustrate the technical solution of the present invention and does not constitute a limitation on the scope of protection of the present invention.

[0027] This invention discloses a liquid level and temperature control system for an industrial chiller, built on a general-purpose industrial controller hardware platform. It connects to a liquid level regulation loop located in the chiller's water tank and a temperature regulation loop located in the circulation pipeline via physical I / O ports. A unidirectional logic channel connecting these two loops is established within the controller. The liquid level regulation loop includes a liquid level sensor for real-time monitoring of the water tank capacity and a water replenishment solenoid valve for performing water replenishment. The temperature regulation loop includes a temperature sensor for collecting circulating medium temperature parameters, an inlet temperature sensor for collecting replenishment medium temperature parameters, and a variable frequency compressor or electronic expansion valve for regulating cooling capacity. This collaborative control device achieves unidirectional transparent information transmission from the liquid level control domain to the temperature control domain through data variable mapping relationships in its internal shared memory area. Addressing the temperature fluctuation problem caused by fluid transport lag and thermal inertia of the temperature sensor during water replenishment in industrial chillers, this system executes a feedforward decoupling control procedure based on enthalpy-current prediction. When the liquid level regulation loop determines that water replenishment is needed based on the liquid level detection value and outputs a water replenishment opening command, the system implements this control. At the same time, the collaborative control device synchronously acquires the temperature parameters of the makeup water medium at the inlet of the makeup water pipeline through the analog input channel within the same millisecond-level scan cycle. and the temperature parameters of the circulating medium in the water tank. The system then calls the flow mapping model pre-stored in non-volatile memory, which contains a data table showing the correspondence between valve opening and mass flow rate. The controller then uses the current... Mass flow rate estimates can be obtained by interpolation from a table. The system executes the logic for calculating the instantaneous enthalpy change rate, and calculates... and The difference, then compare this difference with... and the preset specific heat capacity of the medium Multiplying these together yields the instantaneous enthalpy change rate introduced by the water replenishment action. To convert this thermodynamic parameter into an electrical signal that the controller can execute, the system will... Divide by the compressor's energy efficiency ratio under current operating conditions The equivalent compressor power increment value is obtained as the feedforward compensation signal. The feedback control command superimposed on the output of the PID algorithm in the temperature regulation loop The final power control command is then generated. The variable frequency compressor is driven to change its speed, for example, in a water replenishment event, if calculated... The heat load impact is 2.5kW, and currently... If the value is 3.0, the system will pre-increase the compressor's cooling power by about 0.83kW before the temperature sensor detects a change in water temperature, thereby achieving a pre-balance between energy supply and demand at the source of thermal disturbance.

[0028] Considering that the water replenishment medium is not instantly and uniformly mixed after being injected into the water tank, but requires a certain physical transport time to reach the heat exchange area, this system incorporates a timing shaping module within the collaborative control device to solve the phase synchronization problem between the energy compensation signal and the actual arrival time of the heat cluster. This module operates on a first-order inertial delay function, the transfer function of which is as follows: To address the transmission delay drift caused by flow rate changes in variable frequency water pump applications, this module establishes a communication connection with the circulating pump drive unit to read the pump operating frequency parameters, which characterize the flow rate of the circulating medium, in real time. The dynamic delay mapping logic is based on the formula Real-time calculation of time constant, where To determine the pipeline transport characteristic coefficients obtained through tracer experiments, for example when When the calibration is 150 and the current pump frequency is 50Hz, the time constant It is assigned a value of 3 seconds; if the pump frequency drops to 25Hz, then The automatic adjustment time is 6 seconds. Before being superimposed on the power control command, the feedforward compensation signal undergoes processing by this inertial link, ensuring that the power regulation response curve ultimately applied to the heat exchange actuator closely matches the actual thermal diffusion response curve of the makeup water medium in the water tank in the time domain. This avoids artificial temperature fluctuations caused by premature or delayed compensation actions. To ensure that the time domain response of the feedforward compensation signal is accurately synchronized with the physical transport process of the makeup water medium, the coordinating control device initiates a thermal response hysteresis calibration program to obtain the pipeline transport characteristic coefficients. The calibration procedure is as follows: shut down the heat exchange actuators and the circulating pump. Running at full speed, water tank temperature After stabilization, the trigger duration is as follows: A pulse-like hydration action lasting seconds, recording the time from the start of the hydration action. The temperature of the circulating medium was detected by the sensor. When temperature changes occur Time difference ,calculate Initial calibration values: It is written to a non-volatile memory area and used as a computational benchmark for dynamic delay mapping logic.

[0029] Flow gain coefficient in flow mapping model The initial value is set to The corrected upper and lower boundary values ​​are set as follows: Covering water supply pressure To limit the engineering fluctuation range and prevent coefficient divergence due to sensor malfunctions, the step size is corrected. Set as Deviation integral evaluation time window Set as Instant The time window length is greater than the longest physical mixing time required for the cryogenic medium to fully mix and reach the sensor, ensuring that the integral value reflects the final thermal disturbance residual introduced by the water replenishment action, and triggering the integration threshold of the correction logic. Through calculation Confirmed, among which This represents the standard deviation of the circulating medium temperature sensor during steady-state operation. Exceed When the system determines that the flow model is mismatched, it initiates a step-by-step correction. To prevent system malfunction due to the feedforward compensation exceeding the physical limits of the actuator under extreme high temperatures or full load conditions, the collaborative control device executes disturbance shaping logic based on thermal margin. Before the water supply opening command is generated in the liquid level loop but not yet sent to the actuator, the system reads the current operating status of the inverter in the temperature regulation loop and calculates the remaining power regulation margin. That is, the difference between the maximum rated power and the current output power, based on the current water supply temperature difference, Reverse conversion to the maximum allowable water replenishment flow threshold of the system at this moment The calculation formula follows the principle of energy conservation. The system combines the planned flow rate corresponding to the original planned water replenishment command with... The comparison is performed: if the planned flow rate is less than the threshold, the original instruction is executed; if the planned flow rate exceeds the threshold, the water replenishment opening instruction is limited, reducing the valve opening to extend the water replenishment time, ensuring that the instantaneous enthalpy change rate introduced at any time is always within the range of the compressor's physical regulation capacity, achieving conditional optimal control. The coordinating control device executes disturbance shaping logic based on thermal margin to obtain the remaining power regulation margin of the temperature regulation loop. The calculation is as follows: the rated maximum power of the variable frequency compressor. Subtract the current power output value ,get This power margin The maximum allowable water replenishment mass flow rate threshold for the system at the current moment is calculated by inversely converting the energy conservation relationship. The calculation formula is: ,in The specific heat capacity of the preset medium is used. Temperature of the makeup water medium With circulating medium temperature The difference.

[0030] If the planned flow rate output by the liquid level control circuit exceeds The water supply opening command is limited to a certain limit, restricting the opening of the water supply valve to a certain value. The corresponding opening value, instantaneous enthalpy change rate Strictly constrained Within a certain range, to prevent the compressor from reaching physical saturation under extreme operating conditions, which could lead to temperature runaway, a flow imaging model is used. use Nodes ( to Opening, with A linear interpolation lookup table structure (with step size) is used to store valve openings. Mass flow rate measured under standard water supply pressure Correspondence, estimated actual mass flow rate By outputting values ​​from the lookup table Multiply by the corrected flow gain factor Obtain, that is To address the discrepancy between actual flow rate and model estimates caused by fluctuations in industrial water supply pressure and valve wear, the system incorporates self-calibration logic for model parameters. Within a preset evaluation time window after each water replenishment operation, such as 30 seconds after the water replenishment valve is closed, the system continuously calculates the circulating medium temperature parameters. Relative to temperature setpoint deviation And calculate the integral value. The logic jointly determines the appropriate response based on the polarity of the integral value and the polarity of the water replenishment temperature difference: if cold water is added and the integral value is positive, it indicates that the temperature is too high, meaning the cold energy provided by the feedforward is insufficient. This leads to the conclusion that the actual water replenishment flow rate is greater than the model estimate, thus classifying it as an undercompensated state. Conversely, if the integral value is negative, it is classified as an overcompensated state. Based on the determination result, the system adjusts the flow gain coefficient in the flow mapping model. Perform step-by-step corrections, for example The revised The system will participate in the mass flow estimation for the next water replenishment cycle, thereby enhancing the control system's inherent adaptive capability against pipeline pressure fluctuations and component aging without the need to add flow meters. Furthermore, to address the response lag caused by compressor mechanical inertia, the collaborative control device also executes trend prediction and buffering logic. The system performs differential calculations on the liquid level detection value at 100ms intervals to determine the liquid level descent rate. Based on this, the system calculates the remaining time before the water replenishment threshold is reached. When the remaining time is less than the preset threshold, such as 5 seconds, the system compares the temperatures of the replenishment medium and the circulating medium to determine the polarity of the expected thermal shock. If a thermal shock is determined to occur, the system sends a pre-bias signal to the temperature regulation loop, driving the heat exchange actuator to pre-adjust the circulating medium temperature by 0.1 to 0.2 degrees Celsius below the set value. This creates a thermal potential energy buffer within the water tank to counteract the thermal shock, utilizing volumetric thermal inertia to offset the peak heat load at the initial stage of water replenishment. To eliminate the impact of turbulent noise on the liquid level drop rate caused by the circulating pump suction and return water impact within the storage tank... To mitigate computational interference, the system does not directly perform differential calculations on single-point liquid level samples. Instead, it employs a sliding window least squares fitting algorithm to extract the liquid level change trend. The system maintains a length of [missing information] in memory. like The first-in-first-out (FIFO) data queue is used to store the liquid level sampling sequence within the last 5 seconds in real time. In each calculation cycle, the controller performs linear regression analysis on the position data in the queue relative to the time axis, calculates the slope of the fitted line, and defines it as the current effective liquid level descent rate. The system only considers the goodness-of-fit coefficient. Only when the confidence level exceeds a preset confidence threshold, such as 0.95, is the downward trend of the liquid level confirmed as valid and the remaining time estimate is updated accordingly, thereby filtering out the risk of the pre-bias action being mistakenly triggered by occasional fluctuations in the liquid level.

[0031] Example 1: This example demonstrates the specific operation of a high-power laser cutting equipment cooling system. In this scenario, an industrial chiller provides constant-temperature cooling water to a 12kW fiber laser, and the equipment is operating at a non-full load. To reduce energy consumption, the circulating pump drive unit adjusts the pump operating frequency parameters. The frequency was reduced to 30Hz, causing the flow rate of the circulating medium in the pipeline to fall below the rated flow rate. At this point, due to processing losses, the water level in the tank dropped to the water replenishment threshold, while the temperature parameters of the water replenishment medium in the external water replenishment pipeline... The temperature is 5 degrees Celsius, which is far lower than the temperature parameter of the circulating medium in the water tank. With a set temperature of 25 degrees Celsius, the system faces the control challenge of introducing high-intensity cold shocks under low flow rate conditions. When the liquid level regulation loop detects that the liquid level is too low and generates a water replenishment opening command of 50%, At that time, the collaborative control device initiates the thermodynamic prediction program through a unidirectional logic channel. Within the same logic cycle after the command is issued, the device reads... and The value is obtained, and the mass flow rate estimate corresponding to the current 50% opening is retrieved by calling the flow mapping model. Based on the above parameters, the system calculates the instantaneous enthalpy change rate that the water replenishment action will introduce into the system. And combined with the current compressor's energy efficiency ratio coefficient The feedforward compensation signal that requires the compressor to reduce its output power to balance the cold shock was calculated. This process is completed before the physical medium has been mixed and before the temperature sensor has detected any temperature fluctuations, establishing a logical hard link for hydrothermal action.

[0032] To address the issue of prolonged transport delays caused by current low flow rate conditions, the timing shaping module within the system uses real-time acquired data... The temporal characteristics of the feedforward action are reconstructed, and the system reads the current... Only 30Hz, based on preset dynamic delay mapping logic and pipeline transport characteristic coefficients. The time constant of the longer first-order inertial delay function was calculated. Feedforward compensation signal After processing by the inertial element with a large time constant, the effective curve is lengthened and smoothed on the time axis, so that the compressor power adjustment is no longer a step-like abrupt change, but a slow following process. This precisely synchronizes the actual physical process of the low-temperature makeup water medium mass being transported from the makeup water point and diffused into the heat exchange area in the time domain, avoiding temperature reverse overshoot caused by premature compensation action due to slow flow rate. The system executes disturbance shaping logic based on thermal margin to ensure the safety of the adjustment. Before makeup water is applied, the co-control device reads the current unloading capacity margin of the variable frequency compressor. If the calculation finds that even if the compressor is reduced to its lowest speed, it cannot completely offset the calculated unloading capacity margin... If there is a risk of overcooling, the system will calculate the maximum allowable flow threshold in reverse based on the principle of energy conservation. The system will limit the water replenishment opening command from the originally planned 50% to... The corresponding opening degree is 30%. By actively extending the water replenishment time, the instantaneous heat load intensity is reduced, ensuring that the introduced cold shock is always within the linear range of the compressor's adjustment capability.

[0033] During continuous operation under the aforementioned low flow rate conditions, the system further detected an impending water replenishment event. When the remaining time for triggering the water replenishment action, indicated by the fitted rate of decline of the liquid level detection value, was less than 5 seconds, the trend prediction logic determined that the system was about to suffer a negative thermal shock, given the 20-degree Celsius temperature difference between the replenishment medium and the circulating medium. Within the 3-second window before the water replenishment valve opened, the coordinated control device preemptively sent a signal to the temperature regulation loop. The pre-bias signal drives the compressor to slightly increase its power before the water replenishment action, pre-lowering the water temperature in the tank to 19.8 degrees Celsius. When the replenishment heat mass actually arrives, this pre-stored cold potential energy is instantly neutralized by the thermal shock, so that the final water temperature peak only rises to 20.05 degrees Celsius. This effectively utilizes the thermal capacity inertia of the water body to achieve imperceptible water replenishment. Under the synergistic effect of the above multiple logics, even with the introduction of a replenishment medium with a huge temperature difference and a low circulation velocity, the temperature detection value in the temperature regulation loop is always maintained within a precise range of 0.1 degrees Celsius above and below the set value. This control strategy deterministically correlates the originally separate liquid level action and temperature response through enthalpy-flow calculation, and achieves spatiotemporal synchronization of energy compensation and material transport under variable flow rate conditions, solving the problem of temperature oscillation caused by response lag and phase misalignment in traditional feedback control.

[0034] Example 2: This example aims to construct a rigorous control experiment. The experimental platform is based on an industrial chiller unit with a rated cooling capacity of 20kW and an effective water tank volume of 200L. It is equipped with a variable frequency compressor and a variable frequency circulating pump. To simulate the dynamic thermal characteristics of loads such as lasers, an array of electrically adjustable heating loads is connected to the experimental circuit. The water supply system is connected to the municipal water supply network, and a programmable pressure regulating valve is used to simulate random fluctuations in water supply pressure. The data acquisition system includes high-precision temperature sensors (accuracy ±0.05°C) installed at the water inlet, water tank, and outlet. The present invention includes a sampling frequency of 100Hz and an electromagnetic flowmeter for monitoring circulation flow and replenishment flow. To comprehensively evaluate the superiority of the present invention over the prior art, two sets of control experiments were designed: the control group adopts the traditional independent PID control strategy, in which the liquid level loop is only responsible for switching the replenishment valve, and the temperature loop adjusts the compressor power according to the water tank temperature feedback, with no information interaction between the two; the present invention group adopts the liquid level and temperature coordinated control strategy of the present invention, which includes core functional modules such as feedforward decoupling, timing shaping and model parameter self-calibration. The test process is divided into three stages, which are tested for different types of interference.

[0035] Phase 1: Water Replenishment Thermal Shock Response Test. When the system is in steady-state operation (set water temperature 20°C, load 10kW), a water replenishment action equivalent to 10% of the tank volume is triggered. The temperature of the replenishment medium... Setting the temperature to 10°C, the resulting cold shock, in the control group, resulted in the temperature sensor detecting a temperature drop approximately 15 seconds after the injection of low-temperature water. The PID controller responded immediately, but due to this response lag, the water temperature experienced a drop of approximately 0.8°C and underwent several oscillations before stabilizing. In the present invention's sample group, the collaborative control device calculated the instantaneous enthalpy change rate the moment the water replenishment command was issued. And generate feedforward compensation signal By unloading the compressor in advance, actual test data shows that water temperature fluctuations are limited to within 0.15°C, and there is no obvious oscillation.

[0036] Phase Two: Water Supply Pressure Fluctuation Interference Test. While maintaining water supply, a sinusoidal pressure fluctuation with an amplitude of ±20% and a frequency of 0.1Hz was introduced into the water supply pipeline through a pressure regulating valve. This operation aims to simulate the deviation in water supply flow caused by water pressure changes in actual working conditions, thereby verifying the effectiveness of the model parameter self-calibration logic. The control group, lacking the ability to perceive flow changes and with lag in PID control, experienced continuous periodic deviations in water temperature due to pressure fluctuations, with a deviation amplitude reaching ±0.45°C. The prototype of this invention automatically activated the self-calibration logic after detecting an abnormality in the temperature residual integral, correcting the flow gain coefficient. After two water replenishment cycles of adaptive adjustment, the water temperature fluctuation quickly converged to within ±0.12°C, proving the system's strong stability against pressure disturbances under flow meter-less conditions.

[0037] Phase 3: Adaptability Test for Variable Flow Rate Conditions - This phase will test the frequency of the circulating pump. The Hz frequency was reduced from 50Hz to 30Hz to simulate low-flow-rate conditions under energy-saving operation. This operation altered the transport delay time of the hot spot in the pipeline. The control group did not compensate for the flow rate change, resulting in a mismatch between the originally tuned PID parameters at low flow rates, leading to a slower system response and increased overshoot. The timing shaping module of this invention reads the data in real time. It automatically adjusts the time constant of the first-order inertial delay function based on the dynamic delay mapping logic. The timing of the feedforward signal's activation and the arrival time of the hot spot are always precisely synchronized to ensure the consistency of control performance under variable flow rate conditions. Table 1 summarizes the key performance index data of the above experiments.

[0038] Table 1: Performance Comparison Data Between the Control Sample Group and the Sample Group of the Present Invention

[0039]

[0040] The above test data clearly show that the solution of the present invention exhibits superior performance compared to traditional PID control when dealing with water replenishment thermal shock, water supply pressure fluctuations, and variable flow rate conditions. In particular, through the synergistic effect of feedforward decoupling and adaptive calibration, the system successfully minimizes the impact of various physical disturbances on temperature.

[0041] Example 3: This example combines Figures 1 to 3 A description of a liquid level and temperature control system for an industrial chiller, such as... Figure 1 As shown, the water supply opening command output from the liquid level regulation circuit is transmitted in two paths. One path, after being processed by disturbance shaping logic based on thermal adequacy limiting, generates a limited opening to drive the water supply solenoid valve to perform an action to prevent actuator overload. The other path is synchronously input to the flow image model of the collaborative control device to obtain the estimated mass flow rate from a table. The estimated mass flow rate, combined with the water supply medium temperature collected by the inlet temperature sensor and the current water temperature in the tank collected by the circulating medium temperature sensor, is entered into the instantaneous enthalpy change rate calculation module to calculate the thermal shock introduced by the water supply. This signal is then converted into a feedforward compensation signal, which enters the timing shaping module. This module combines the flow rate parameters and pump operating frequency parameters obtained from the circulating pump drive unit, and uses a first-order inertial delay function to process the signal to dynamically synchronize the arrival time of the hot spot. The generated timing-synchronized power control command is superimposed on the temperature regulation loop to drive the variable frequency compressor or expansion valve, ultimately generating a temperature response residual. The system calculates and outputs the corrected flow gain coefficient based on the temperature response residual and the self-calibration logic of the operating model parameters. Feedback is fed back to the flow mapping model, thereby forming a closed-loop adaptive mechanism to combat water pressure fluctuations and aging.

[0042] like Figure 2 As shown in the figure, the horizontal axis represents time in seconds, the left vertical axis represents temperature in degrees Celsius (°C), and the right vertical axis represents the normalized pressure fluctuation amplitude. The dashed line indicates the oscillations and deviations in temperature caused by independent PID control under pressure fluctuation interference, while the solid line indicates that the temperature of the system of this invention remains stable near the set value of 20.0°C under the same water supply pressure fluctuation interference. Figure 3 As shown, the hardware architecture of this system is centered on a central control unit programmable controller or embedded system that integrates a collaborative control program including disturbance shaping, timing synchronization, flow mapping database, and a background model parameter self-calibration module. A logical channel is established through an internal shared data area. The input end is connected to an inlet temperature sensor for collecting simulated signals from the water supply pipeline, a circulating medium temperature sensor for collecting simulated signals from the outlet or water tank, a liquid level monitoring device for collecting water tank level signals, and a circulating pump driver frequency converter that provides feedback frequency via digital communication. The output end is connected to a controlled water supply solenoid valve to receive opening control commands, and a variable frequency compressor unit to receive power adjustment commands. All actuators act on the physical controlled object, namely the industrial chiller water circulation system, which includes the liquid storage tank.

[0043] Example 4: This example aims to deeply explain the protection mechanism and execution process of the disturbance shaping logic based on thermal margin when the refrigeration compressor approaches the physical performance boundary. The disturbance shaping logic is described as a reverse capability constraint mechanism. This example will reveal how this logic transforms the abstract power margin into a specific flow limiting action through a specific full-load operating condition simulation, ensuring the safety of the system under extreme conditions. The scenario is set as an industrial chiller operating at full load in a high-temperature environment in summer (ambient temperature 40°C). The compressor is already at 95% of its rated speed, i.e., a load rate of 95%. At this time, the water tank level triggers a water replenishment request, and the temperature of the replenishment medium is... The target temperature for the circulating medium is 35°C. At 20°C, the expected thermal shock introduced by the water replenishment operation is calculated as follows: if water is replenished at full speed (water replenishment valve opening 100%) according to conventional logic, the instantaneous enthalpy change rate is... This will require the compressor to increase its power output by an additional 15%, and the compressor's physical residual power adjustment margin... With only 5% (100%-95%), forcibly implementing full-speed water replenishment will inevitably lead to compressor overload and saturation, causing the water temperature to rise uncontrollably.

[0044] After receiving the 100% water replenishment opening command from the liquid level circuit, the collaborative control device performs capacity boundary calculations, and the system reads the current compressor load rate in real time. and rated maximum power The remaining available power was calculated. Based on the current water replenishment temperature difference and specific heat capacity of the medium The maximum mass flow rate threshold that this power margin can withstand is calculated in reverse. The calculation formula is: Assuming calculation Corresponding to 30% opening of the water supply valve; the system performs action shaping, comparing the 100% opening command requested by the liquid level circuit with the calculated 30% limit value, taking the smaller value as the final execution command, and the coordinating control device outputs a 30% opening command to drive the water supply valve. Although the water supply flow is limited, resulting in the water supply process taking about 3 times longer, the introduced instantaneous thermal shock is strictly limited to the compressor's remaining 5% adjustment capacity. The compressor then smoothly accelerates to 100% full load, successfully offsetting the thermal disturbance after the limit, ensuring that the water temperature fluctuation is always under control throughout the water supply process, and no over-temperature alarm or shutdown accident occurs. This process demonstrates how the system sacrifices the time dimension (water supply rate) to achieve a safe balance in the energy dimension (cooling capacity).

[0045] Example 5: This example aims to deeply analyze and verify the self-calibration logic of the model parameters and the decision rules behind it. The test object is the control system of the aforementioned industrial chiller, and the core task is to measure the external water supply pressure. When unexpected fluctuations occur, the flow gain coefficient in the flow mapping model is adaptively adjusted. To maintain the accuracy of feedforward control, in the initial state, the system's preset flow mapping model is based on the standard water supply pressure. Calibration, at this time In actual industrial settings, water supply pressure often experiences random disturbances. This embodiment constructs a pressure disturbance simulation platform, introducing sinusoidal pressure fluctuations into the water supply pipeline via a high-frequency proportional pressure reducing valve. The fluctuation range is set to... to The frequency is To simulate typical municipal pipeline network pressure instability; the core decision-making logic chain for parameter self-calibration is as follows: the system sets the temperature deviation integral threshold. The threshold is determined based on the noise level of the temperature sensor. and the allowable steady-state temperature error band Taking all factors into consideration, to avoid sensor random noise falsely triggering the calibration action, Must meet ,in To evaluate the length of the time window, in this embodiment, it was determined that... ,set up After calculation, take After the hydration process is completed Absolute value of the integral of the temperature deviation within When the system determines that the flow model is mismatched, it initiates parameter correction and adjusts the step size. The settings follow the principle of small steps and rapid iterations, aiming to ensure convergence speed while preventing overshoot oscillations. The optimal correction step size is determined through offline simulation and on-site debugging. .

[0046] During the test, the water supply pressure was adjusted stepwise to... At this point, for the same valve opening command, the actual water supply flow rate is greater than the model estimate, resulting in insufficient feedforward cooling capacity compensation. After the first water supply is completed, the system calculates the integral value of the temperature deviation. Far exceeding the threshold Based on the calibration logic, the system is determined to be in an undercompensated state and will automatically execute [the necessary steps]. During the subsequent second hydration maneuver, although the pressure remained at However, due to As the value increases, the feedforward compensation increases accordingly, and the integral value of the temperature deviation converges to... If the value is less than the threshold, the system determines that the model has been temporarily matched and stops the correction. This process clearly demonstrates how the parameter self-calibration logic achieves indirect perception and compensation for unmeasurable pressure disturbances through temperature residual feedback. Table 2 records the parameter evolution process data under this pressure step disturbance.

[0047] Table 2: Data Recording Table of Parameter Self-calibration Process under Pressure Step Disturbance

[0048]

[0049] The above data and logical deductions show that the model parameter self-calibration mechanism based on thermal response residuals proposed in this invention can adaptively track and compensate for changes in external environmental parameters without the assistance of a flow meter, based on clear quantization thresholds and step rules.

[0050] Example 6: This example aims to provide a standardized offline calibration and data filling procedure, specifically for the core technical component of the flow mapping model, which is essentially based on valve opening. To eliminate the black-box nature of the nonlinear lookup table function in this model, this procedure constructs a standard hydraulic test bench containing a high-precision mass flow meter. A full-stroke characteristic scan is performed on each type of proposed water supply actuator. During the test, the valve opening is increased in 1% increments from 0% to 100%, and the system operates stably for at least 10 seconds at each opening point. The corresponding steady-state mass flow rate value is recorded. The least squares method is used to perform polynomial fitting on the discrete measurement points, generating a continuous flow characteristic curve. A lookup table containing 101 nodes is stored in the controller's non-volatile memory area, ensuring that the data source of the flow mapping model has traceable physical measurement evidence, rather than relying on theoretical assumptions.

[0051] This embodiment also provides a set of on-site pre-deployment calibration procedures. When the control system is first deployed at a specific industrial chiller site, the system automatically runs the initialization calibration program. During the initial water filling process, the water supply valve opening is manually or automatically fixed at 50%, and the time required for the liquid level to rise to a specific height is recorded using a stopwatch or auxiliary metering tool. This allows the actual average flow rate to be calculated. The control system then compares this measured flow rate with the theoretical flow rate corresponding to 50% opening in the flow mapping model to calculate the initial flow correction coefficient. This coefficient serves as the initial reference for system operation, replacing the default value of 1.0. This eliminates the systematic deviation caused by the inconsistency between the on-site water supply pressure and the calibrated pressure from the very beginning of system operation, ensuring that the feedforward control logic has high-precision adjustment capabilities from the first day.

[0052] Example 7: This example aims to provide a standardized system deployment pre-deployment calibration and engineering commissioning procedure to adapt to refrigeration hardware produced in different batches and varying on-site heat load characteristics. It addresses key parameters in the aforementioned timing shaping module, including pipeline transport characteristic coefficients. This coefficient directly determines the synchronization accuracy between the feedforward signal and the arrival time of the hot spot. Given that this coefficient is highly dependent on the physical layout and pipe diameter resistance characteristics of the specific pipeline, it is impossible to cover all application scenarios with a single fixed value preset at the factory. Therefore, this procedure requires that an automated thermal response hysteresis calibration program be executed after the equipment is installed on-site and the initial water filling is completed. During program execution, the control system shuts down the compressor and starts the circulating pump at full speed. Once the water temperature stabilizes, a pulse-like water replenishment action is triggered, such as opening the water replenishment valve for 5 seconds. The system continuously monitors the data from the water tank temperature sensor and records the data from the start of the water replenishment action. Until the temperature reading begins to decrease Time difference between According to the physical definition, the pipeline transport characteristic coefficient at this time can be calibrated as follows: This calibration value will be automatically written to the controller's non-volatile memory area as a calculation benchmark for subsequent dynamic delay mapping logic, thereby eliminating model errors introduced by differences in pipeline length.

[0053] Furthermore, to address the differences in the dynamic characteristics of heat source loads under various application scenarios, this embodiment also includes an adaptive PID parameter tuning procedure. When the system is running for the first time or when there is a significant change in load characteristics, it automatically enters a relay feedback self-tuning mode. In this mode, the controller temporarily bypasses the feedforward channel and uses the feedback loop to perform on-off control of the set temperature, inducing the system to generate constant-amplitude oscillations. The system records the critical period of the temperature oscillations in real time. and critical gain And based on the Ziegler-Nichols rule or other improved engineering tuning formulas, the optimal proportional coefficient is calculated. Integral Time and differential time After the tuning is completed, the system automatically switches back to the normal cooperative control mode. This procedure ensures that the feedback loop can provide the best steady-state accuracy and anti-disturbance capability under load conditions with different heat capacity and heat dissipation characteristics.

[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 present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0055] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A liquid level and temperature control system for an industrial chiller, comprising a liquid level regulating loop, a temperature regulating loop, and a collaborative control device, wherein the collaborative control device establishes a unidirectional logic channel connecting the liquid level regulating loop and the temperature regulating loop, characterized in that, The collaborative control device executes the following control logic: Within the synchronous clock cycle of the water replenishment opening command output by the liquid level regulation loop, the temperature parameters of the water replenishment medium and the temperature parameters of the circulating medium are obtained. The flow image model pre-installed in the cooperative control device is called to find the mass flow rate estimate corresponding to the water replenishment opening command. Based on the mass flow rate estimate, the instantaneous enthalpy change rate introduced by the water replenishment action is calculated. The instantaneous enthalpy change rate is converted into a feedforward compensation signal with the same dimensions as the temperature regulation loop and superimposed on the power control command of the temperature regulation loop. The collaborative control device also executes model parameter self-calibration logic: within the preset evaluation time window after the water replenishment action corresponding to the water replenishment opening command ends, it calculates the integral of the deviation of the circulating medium temperature parameter relative to the temperature setpoint. Based on the polarity of the deviation integral and the polarity of the difference between the temperature parameters of the makeup water medium and the temperature parameters of the circulating medium, it is determined whether the feedforward compensation of the previous makeup water cycle is under-compensated or over-compensated. Based on the judgment results, the flow gain coefficient in the flow image model is modified step by step, and the modified flow gain coefficient is used to update the flow image model to participate in the mass flow estimation of the next water replenishment cycle, so as to offset the model error introduced by the pressure fluctuation of the water replenishment pipeline or the drift of the actuator characteristics. The calculation logic for the instantaneous enthalpy change rate is as follows: calculate the temperature difference between the temperature parameters of the makeup water medium and the temperature parameters of the circulating medium, calculate the product of the temperature difference, the estimated mass flow rate, and the preset specific heat capacity of the medium, and define this product as the instantaneous enthalpy change rate; the conversion logic for the feedforward compensation signal is as follows: divide the instantaneous enthalpy change rate by the energy efficiency ratio coefficient of the compressor in the temperature regulation loop to obtain the equivalent compressor power increment value as the feedforward compensation signal.

2. The liquid level and temperature control system for an industrial chiller according to claim 1, characterized in that, The collaborative control device also includes a timing shaping module, which operates a first-order inertial delay function. The collaborative control device uses the first-order inertial delay function to process the feedforward compensation signal, and then superimposes the processed feedforward compensation signal onto the power control command, so that the power regulation response curve of the heat exchange actuator controlled by the temperature regulation loop is synchronized with the thermal diffusion response curve of the makeup water medium in the chiller tank in the time domain.

3. The liquid level and temperature control system for an industrial chiller according to claim 2, characterized in that, The collaborative control device also establishes a communication connection with the circulating pump drive unit to obtain pump operating frequency parameters characterizing the flow velocity of the circulating medium; the timing shaping module operates dynamic delay mapping logic, which calculates the time constant of the first-order inertial delay function in real time according to the following formula: ,in, Let be the time constant of the first-order inertial delay function. The preset pipeline transport characteristic coefficient, The pump operating frequency parameter is used; the timing shaping module dynamically assigns the time constant to the calculation result, so that the effective time of the feedforward compensation signal automatically drifts with the change of the circulating medium flow rate.

4. The liquid level and temperature control system for an industrial chiller according to claim 1, characterized in that, The collaborative control device also executes disturbance shaping logic based on thermal margin: before responding to the water supply opening command, it obtains the current remaining power regulation margin of the temperature regulation loop; based on the temperature difference between the water supply medium temperature parameter and the circulating medium temperature parameter, it reversely converts the remaining power regulation margin into the maximum water supply flow threshold that the system can introduce at the current moment. The planned flow rate mapped from the water replenishment opening command is compared with the maximum water replenishment flow rate threshold. If the planned flow rate exceeds the maximum water replenishment flow rate threshold, the water replenishment opening command is limited to ensure that the instantaneous enthalpy change rate introduced by the water replenishment action is always within the physical regulation capability range of the temperature regulation loop.

5. The liquid level and temperature control system for an industrial chiller according to claim 1, characterized in that, The collaborative control device also performs trend prediction and buffering logic: it monitors the rate of decrease of the liquid level detection value in the liquid level regulation loop and estimates the remaining time to trigger the water replenishment action; when the remaining time is less than the preset threshold, it compares the temperature parameters of the water replenishment medium with the temperature parameters of the circulating medium to determine the polarity of the expected thermal shock; before the water replenishment action is triggered, it sends a pre-bias signal to the temperature regulation loop according to the polarity, driving the temperature regulation loop to pre-adjust the temperature parameters of the circulating medium in the opposite direction to the expected thermal shock, thus building a thermal potential energy buffer against thermal shock in the liquid storage tank.

6. The liquid level and temperature control system for an industrial chiller according to claim 1, characterized in that, The logic for determining whether the feedforward compensation in the previous water replenishment cycle is under-compensated or over-compensated includes: if the temperature parameter of the replenishment medium is lower than that of the circulating medium and the deviation integral is positive, it is determined to be under-compensated, indicating that the actual water replenishment flow rate is greater than the estimated mass flow rate; if the temperature parameter of the replenishment medium is lower than that of the circulating medium and the deviation integral is negative, it is determined to be over-compensated, indicating that the actual water replenishment flow rate is less than the estimated mass flow rate.

7. The liquid level and temperature control system for an industrial chiller according to claim 1, characterized in that, The collaborative control device obtains the temperature parameters of the replenishing medium by reading the data from the temperature sensor installed at the inlet of the water replenishment pipeline of the chiller; the collaborative control device obtains the temperature parameters of the circulating medium by reading the data from the temperature sensor installed at the outlet of the chiller; the liquid level regulation circuit includes a liquid level sensor and a water replenishment solenoid valve, and the water replenishment opening command is used to control the duty cycle or opening percentage of the opening time of the water replenishment solenoid valve.

8. The liquid level and temperature control system for an industrial chiller according to claim 1, characterized in that, The logic of step-by-step correction of the flow gain coefficient is as follows: when it is determined to be undercompensated, the preset step value is increased based on the current flow gain coefficient; when it is determined to be overcompensated, the preset step value is decreased based on the current flow gain coefficient. The initial value of the flow gain coefficient is preset to 1.0, and upper and lower limit boundary values ​​that allow for correction are set to prevent the coefficient from diverging due to temperature sensor failure.

9. The liquid level and temperature control system for an industrial chiller according to claim 1, characterized in that, The collaborative control device is integrated into a programmable logic controller or embedded microcontroller. The unidirectional logic channel is the data variable mapping relationship in the shared memory area inside the controller. The power control command of the temperature regulation loop is used to drive the variable frequency compressor or electronic expansion valve to regulate the cooling capacity output of the refrigeration system.

Images