PASSIVE THERMAL ENERGY STORAGE IN REFRIGERATED WAREHOUSES ABSTRACT 1. INTRODUCTION

Transcription

1 PASSIVE THERMAL ENERGY STORAGE IN REFRIGERATED WAREHOUSES J. E. ALTWIES AND D. T. REINDL College of Engineering, University of Wisconsin, Madison, WI, USA ABSTRACT This paper investigates operational strategies that use stored products as thermal mass to shift refrigeration loads to more favorable operational periods (low energy cost periods, lower outdoor air conditions, etc.) allowing an opportunity to reduce system operating costs. An integrated model of the stored product, warehouse air, and warehouse structure is developed and thermal response characteristics are predicted for a selected low temperature warehousing facility. Simulated results are validated with experimental measurements. Food quality impacts associated with the temperature cycling caused by potential operating strategies are discussed. Results from this investigation indicated that a full load-shifting control strategy would save $82,000 (US) ($0.40/ft 2 /yr or $4.28/m 2 /yr) annually over the test facility s current operational strategy, representing 53% of the total cooling cost. Predicted maximum warehouse temperature variation is 5.6 C (10 F), which is not expected to cause significant product quality changes in the temperature range (below 18 C) studied. 1. INTRODUCTION In the U.S., the industrial sector is responsible for the largest aggregate consumption of electricity (EIA, 1998). Industrial refrigeration applications, specifically refrigerated warehouses, are the focus of this paper. Industrial refrigeration systems are significant energy consumers. These systems usually operate during the daytime when electrical costs and outdoor temperatures are highest and refrigeration system performance is at its worst. This study was initiated to test the feasibility of using the thermal mass of the items in storage as a means of decoupling the operation of the refrigeration system from the loads that it serves. In this case, refrigeration equipment operates during low utility cost hours (off-peak) to pre-cool the stored items. Then the refrigeration equipment can remain idle during high utility cost periods (on-peak) with minimal changes in the storage environment and product temperature. In many cases, little or no capital investment is required to implement this type of warehouse operating strategy. Tests and analyses were conducted at a low temperature facility located in the north central U.S. Four individual warehouses, labeled A through D, are collectively operated at this site, which processes, packages, and stores frozen vegetables. The warehouses have an aggregate product (frozen vegetables) storage capacity that exceeds 22 million kg (50 million lb) at full capacity. 2. FOOD QUALITY ISSUES The advantages and benefits of demand shifting are cost savings for the customer and reduced on-peak demand for the electric utility. However, possible disadvantages include risk to the product quality, shelf life, and nutrient content. While the proposed operating strategies are designed to minimize changes in the storage environment, even minor alterations and fluctuations merit investigation of possible detrimental effects on product quality. Energy and cost savings are of no use if they lead to degradation in product quality. Time constraints for this project did not allow for a direct investigation of food quality effects; however, an in-depth survey of previous investigations was conducted and their results compiled. Studies by Aparicio-Cuesta and Garcia-Moreno (1988), Ashby et al. (1979), Boggs et al. (1960), Gortner et al. (1948), Hustrulid and Winter (1943), and Woodroof and Shelor (1947) involved a variety of food products (spinach, cauliflower, peas, strawberries, and others) and were undertaken for reasons ranging from food quality deterioration to energy savings. Each study investigated the 20 th International Congress of Refrigeration, IIR / IIF, Sydney,

2 effect of temperature fluctuations over ranges of interest for each particular project. The studies that never allowed frozen products to exceed 0 C (32 F) consistently obtained results showing little to no degradation attributable to fluctuations in the storage temperature. The studies allowing extreme temperature fluctuations with freeze / thaw cycles reported significant quality losses. We show that the product temperature fluctuations expected from a full demand shifting strategy are comparable to the study by Ashby et al. (1979) and product temperatures are required to stay below 18 C (0 F). For the refrigeration system operating strategies considered in this investigation, nutrient retention is expected to remain the same or even improve, due to the decrease in average storage temperature from the set point of 20.6 C ( 5 F) to 21.9 C ( 7.5 F). The research by Ashby et al. (1979) also indicates that the periodic temperature variations that accompany the methods explained in this paper will not affect nutrient loss. 3. WAREHOUSE COOLING LOAD The cooling load in the warehouse must be modeled to calculate the air temperature rise during periods without active refrigeration system operation (high energy cost or on-peak periods). The total cooling load is the sum of five contributions: transmission load, product load, internal load, infiltration load, and equipment load. Transmission load is defined as the heat gain through the exterior surfaces (e.g. walls and roof, which are adjusted to include the effects of solar radiation absorbed by the surface) and floor (attributable to a glycol heating loop imbedded in the concrete) of a conditioned space. Product load is represented by the rate of energy transfer to the refrigerated space due to the entrance of warm products, which are stored in 1.81 m 3 (64 ft 3 ) cardboard boxes. People, lights, forklifts, and other miscellaneous heat-generating objects within the space comprise the internal load. Infiltration is the uncontrolled exchange of non-conditioned air with the conditioned warehouse air through doorways or other openings. This air exchange adds both sensible and latent loads to the space. On-site observations of traffic through each doorway provided estimates of the frequency of door openings and dwell times, which allowed infiltration rates to be developed using the methods documented in ASHRAE (1990). The fifth component of the refrigeration load is attributable to equipment. Any heat added to the space by the refrigeration equipment itself, such as evaporator fans and defrosts, is included in this category. Since the warehouses operate 24 hours a day, hour-to-hour variations in non-transmission loads are minor. Basic information about the construction of the warehouses was obtained using blueprints provided by staff at the facility and is summarized in Table 1. Warehouses A & B were built together during the 1960 s, with identical construction. As a result, a single cooling load model was created for both. However, warehouses C and D were constructed separately within the last 15 years. The building materials used in the walls and roof were different from A and B and from each other, requiring the creation of a separate model for each. Design weather conditions used for the calculations are 37 C/26 C (98 F/78 F). Table 1: Warehouse Information Size, m^2 (ft^2) Vintage U Value -- Walls, W/m^2-C (Btu/hr-ft^2-F) Primary Structural Component Lightweight or Heavyweight Installed Capacity, kw (tons) Operating Set Point, C ( F) Warehouse "A" & "B" Warehouse "C" Warehouse "D" 9204 (99069) 2412 (25963) 7420 (79869) (0.036) (0.046) (0.046) Wood / Insulation Concrete / Insulation Concrete / Insulation Lightweight 281 (80) (-5) Heavyweight 422 (120) (-5) Heavyweight 1122 (319) (-5) 20 th International Congress of Refrigeration, IIR / IIF, Sydney,

3 All necessary information about the operation of the equipment, facility construction, and schedules were supplied by the facility staff and used to calculate this portion of the total cooling load. Figure 1 shows the hourly variation of the calculated design cooling load for each of the four warehouses Hour On Peak Period 12-Hour On Peak Period 169 Load (kw) for "A", "B", & "D" Warehouse "C" Warehouse "D" Warehouses "A" & "B" Load (kw) for "C" Hour of Day Figure 1: Total Refrigeration Load: Design Conditions 4. WAREHOUSE MODEL A finite element model of the warehouse structure, air volume within the warehouse, and stored product was constructed to predict distributed temperatures within the stored product and the warehouse air temperature. The entire warehouse model is carefully proportioned to match the actual dimensions of the facility, including the product-to-air volume ratio and the wall layer thicknesses. The previously described loads are applied to the finite element warehouse model by separating the transmission load from the remainder of the total refrigeration load. The transmission portion of the cooling load must be treated separately, because the actual transmission heat gain rate is dependent on the difference between the variable exterior temperature and the now-changing warehouse air temperature (i.e., the rate of energy flowing through the walls, roof, and floor decreases as the temperature inside the warehouse rises). To accurately represent this phenomenon and model each warehouse as a whole, all non-transmission loads are applied to the model as internal heat generation (per unit volume of air) within the volume of air surrounding the stored product. The exterior surfaces of the warehouse are each modeled with a heat transfer coefficient and surface temperature. The heat transfer coefficient of 17 W/m 2 -K (2.9 Btu/hr-ft 2 -F) is a standard value listed in chapter 28 of the 1997 ASHRAE Fundamentals handbook for exterior convection and long-wave radiation (ASHRAE, 1997). The temperature used is the effective outdoor air temperature (allowing for absorbed solar radiation). A constant temperature boundary condition of 19.7 C (67.5 F) is assumed for the floor, which is the average of the supply and return glycol temperatures for the underfloor heater. Separate models were created for low temperature warehouses A & B, C, and D using this process. Figures 2 and 3 illustrate the temperature rise over a 14-hour on-peak window (8 a.m. to 10 p.m.) during a design day when all refrigeration for the warehouse is idle for the worst case warehouse ( A & B ) and the best case warehouse 20 th International Congress of Refrigeration, IIR / IIF, Sydney,

4 ( D ), respectively. The results for warehouse "C" are not shown simply due to space constraints for this paper. The temperatures at three points within the warehouse are shown: air, a corner of the product block (represents the highest product temperature), and the center of the product block. Assuming the initial temperature within the warehouse is 20.6 C ( 5 F), the model predicts a final air temperature above the acceptable maximum temperature limit ( 18 C, 0 F) for long-term storage of frozen vegetables. In warehouses A, B, and C, the product corners (worst case product temperature location) also rise above 18 C (0 F), indicating that these warehouses would need to be pre-cooled below 20.6 C ( 5 F) in order to prevent product temperatures from rising above 18 C (0 F) by the end of the on-peak period. Only warehouse D stays below the limit, even on the most exposed areas (corners). Differences in the model for warehouse D, including the construction of the roof and the increased volume of air, contribute to this result [ C] Air temperature -19 Corner temperature Center temperature Figure 2: Predicted Temperature Rise: Warehouses A & B [ C] Air temperature Corner temperature Center temperature Figure 3: Predicted Temperature Rise: Warehouse D To guarantee that the models were predicting the true conditions in the warehouses, temperature measurements were taken at all three warehouses during a variety of weather conditions. These weather conditions were then applied to the model, and minor differences between the measured 20 th International Congress of Refrigeration, IIR / IIF, Sydney,

5 and predicted results were corrected by adjusting the internal load value due to people in the space. The number of people in the warehouses in any given hour was originally estimated, and the amount of error introduced to the model from this estimate was reasonable. With the load value adjusted, the model gave predicted results that matched the experimental data for all measured weather conditions. 5. DEVELOPMENT OF DEMAND-SHIFTING STRATEGIES Three demand-shifting options were analyzed for possible implementation at the facility. The first option, full demand-shifting, requires a complete shutdown of all refrigeration equipment during the specified 14-hour on-peak time period. This option offers the greatest demand reduction leading to the greatest energy cost savings, but it requires the greatest installed refrigeration capacity. In this case, the refrigeration system must be able to extract the entire integrated refrigeration load generated over a cycle, i.e. 24 hours, in only the 10 hours available during the offpeak window. Figure 4, which shows the model-generated temperature variations inside warehouse A & B, is included as an example of the expected temperature response using full demand shifting. The plot displays results for a 72 hour period, corresponding to three consecutive design days. Under this option, warehouse "A" & "B" would need 834 kw (237 tons) of refrigeration capacity. -14 [ C] Corner Air Center Figure 4: Warehouses A & B : Full Shifting Strategy, Three Design Days A second option, load leveling, meets the refrigeration load in each warehouse by running the equipment at a constant level throughout the 24-hour period. In effect, this option slightly overcools the warehouse at night and slightly under-cools it during the day as the warehouse loads vary. This operating strategy offers the least savings of the three options, but requires the least amount of installed equipment capacity. For example, warehouse "A" & "B" would need 369 kw (105 tons) of capacity for this option. Figure 5 is included as an example of the effect of this operating strategy. While the figure shows that the corner temperature is slowly rising after three consecutive days of design weather, the extremely rare occurrence of many sustained days of design weather would be needed to approach the -18 C (0 F) limit. The third option is a combination of the first two. The refrigeration runs at one level during the day and a higher level at night. This demand-limiting strategy requires less total refrigeration system capacity than full demand shifting, but offers more savings than load leveling. The 20 th International Congress of Refrigeration, IIR / IIF, Sydney,

6 additional savings are gained during colder months, when the overall refrigeration load is lower. Depending on the capacity of the warehouse, full demand shifting may be possible during the winter months under this third alternative. During January, warehouse "A" & "B" could run at 369 kw (105 tons) at night, but reduce to 208 kw (59 tons) during peak hours. Under all three operating strategies, the product temperature stays below the 18 C (0 F) limit [ C] Center Air Corner Figure 5: Warehouses A & B : Load Leveling Strategy, Three Design Days 7. ECONOMIC ANALYSIS An economic analysis of each option was conducted. The first step in this process is to analyze a base case option, i.e., direct on-demand refrigeration. The total annual refrigeration load for each warehouse was calculated using an entire year of typical hourly weather data. Next, the total annual cost of operating the warehouses under normal conditions is calculated by knowing the amount of power the refrigeration system uses per unit of cooling. The system performance is dependent, in part, on the outside air wet-bulb temperature, so it changes hourly throughout the year. Additional electrical loads attributable to the warehouses must also be added, such as lights and evaporator fans. In effect, these items cost double to operate, since the initial electrical power must be purchased, then the heat they generate must be removed by the refrigeration system. Once the total electrical load is calculated, the annual cost for normal operation is computed using the facilities utility rate structure. In this case, all weekends, holidays, and weeknights (between 10 PM and the following 8 AM) are considered off-peak times. The next step in the economic analysis is to determine the costs associated with each of the three operating strategies, then calculate the savings over normal operation in each case. The method used to calculate the cost of normal operation is also used for each demand-shifting option, so differences in the total costs are due to variations in refrigeration system operation during the onand off-peak periods. Under normal conditions, the annual cost to operate the ammonia refrigeration equipment that services all warehouses is nearly $155,000 (U.S.). Using full demandshifting, annual operating costs can be reduced to $73,000 (U.S.) - a 53% savings. However, additional evaporator capacity must be added at this particular facility in order to implement this option. Assuming that additional capacity has an associated capital cost of $1,000 (U.S.) per ton ($284/kWr), full demand shifting will require an investment of about $130,000 (U.S.). Combined with the savings this method would generate, the simple return-on-investment is 62%. With loadleveling, only about $1,000 (U.S.) is saved. Using the third option, slightly more than $58,000 (U.S.) is saved and no additional capacity is required. This savings is realized when the night and 20 th International Congress of Refrigeration, IIR / IIF, Sydney,

7 day operating levels are chosen perfectly. In real operation, the savings will likely be smaller due to errors in forecasting refrigeration loads. The economic projections presented here are dependent on a variety of factors. For example, if the estimated cost of adding additional refrigeration capacity is more than $1,000 (U.S.) per ton ($284/kWr), then the cost of upgrading the system for demand shifting will be greater than what is shown. The operating parameters of the refrigeration system, including the minimum allowable condenser temperature and saturated suction temperature, also influence the cost predictions. The values shown here are based on the information supplied by the warehouse operators. 8. CONCLUSIONS / RECOMMENDATIONS The feasibility of developing and implementing a strategy that shifts refrigeration loads by using stored product as thermal mass has been investigated. A large cold storage warehouse in the upper midwest U.S. was used as a basis to evaluate different refrigeration system operating strategies. In this case, full demand shifting is expected to reduce operating costs by 53% leading to $82,000 (U.S.) savings annually. Another strategy considered combines the full demand shifting and load leveling techniques to achieve annual savings of $58,000 (U.S.) (37%) without the need for additional capacity. However, this technique requires a comparatively sophisticated control strategy for actual implementation. Load leveling alone only saves about $1,000 (U.S.) annually. The largest product temperature fluctuations, from C to 19.1 C ( 12.5 F to 2.5 F), occur using the full demand shifting technique. This range of variation is not expected to have any detrimental effects on the quality of the products in storage. Several factors concerning the operation of the refrigeration equipment affect the total load and energy costs. Among these factors are the minimum allowable condenser temperature, saturated suction temperature, and optional fan cycling during on-peak operation. In general, the highest possible suction temperature and the lowest possible condenser temperature are desirable for optimal system performance. Based on the experimental data recorded at the site, periodic fan cycling in each warehouse is recommended. Cycling the evaporator fans for a few minutes each hour does not add any significant cost, and will reduce the effects of stratification. In order to control the cycling of fans, routinely monitor warehouse and refrigerant temperatures, and efficiently manage any operating strategy, the installation of a control system is also recommended. 11. REFERENCES 1. American Society of Heating, Refrigerating, and Air-Conditioning Engineers (ASHRAE), 1997, ASHRAE Handbook Fundamentals (I-P Edition). 2. American Society of Heating, Refrigerating, and Air-Conditioning Engineers (ASHRAE), 1990, ASHRAE Handbook Refrigeration (I-P Edition). 3. Aparicio-Cuesta, M.P. and C. Garcia-Moreno, 1988, Quality of Frozen Cauliflower during Storage, Journal of Food Science, vol. 53 no. 2: p Ashby, B.H., A.H. Bennett, W.A. Bailey, W. Moleeratanond, and A. Kramer, 1979, Energy Savings and Quality Deterioration from Holding Frozen Foods at Two Daily Temperature Levels, Transactions of the ASAE, p Boggs, M.M., W.C. Dietrich, M. Nutting, R.L. Olson, F.E. Lindquist, G.S. Bohart, M.J. Neumann, and M.J. Morris, 1960, Time-Temperature Tolerance of Frozen Foods XXI: Frozen Peas, Food Technology, vol. 14 : p Energy Information Administration, 1998, Energy in the United States: A Brief History and Current Trends: Annual Energy Review Gortner, W.A., F. Fenton, F.E. Volz, and E. Gleim, 1948, Effect of Fluctuating Storage Temperatures on Quality of Frozen Foods, Industrial Engineering and Chemistry, vol. 40, no. 8: p Hustrulid, A. and J.D. Winter, 1943, The Effect of Fluctuating Storage Temperature on Frozen Fruits and Vegetables, Agricultural Engineering, vol. 24, no. 12: p Woodroof, J.G. and E. Shelor, 1947, Effect of Freezing Storage on Strawberries, Blackberries, Raspberries, and Peaches, Food Freezing, February: p , th International Congress of Refrigeration, IIR / IIF, Sydney,

8 LE STOCKAGE PASSIF D ENÉRGIE THERMIQUE DANS DES ENTREPOTS RÉFRIGERES RÉSUMÉ: Cet article étudie des stratégies opérationelles qui emploient un produit alimentaire comme masse thérmique pour déplacer la consomation de refroidissement aux périodes d opérations plus favorables (des coûts d énérgie plus bas, des conditions environmentales plus basses, etc.) ce qui permet de réduire les coûts de l opération du système. Un modèle intégré du produit, de l air d entrepôt, et de la structure d entrepôt est développé et les caractéristiques de la réponse thermique sont simulés pour un entrepôt choisi. Les résultats de la simulation sont validés par des experiences sur place. Des conséquences sur la qualité des produits alimentaires associés aux cycles de la température provoquées par des stratégies de contrôle proposées sont étudiés. Les résultats de cette investigation indiquent qu une stratégie qui déplace la consommation entiere pourrait économiser $82,000 (US) ($4.28/m 2 /yr ou $0.40/ft 2 /yr) chaque année au-dessus de la stratégie actuelle d entrepôt; un chiffre qui représente 53% du coût de refroidissement. La variation maximum de température prévue pour l entrepôt (5.6 C ou 10 F) n est pas suffisant pour abaisser la qualité du produit de manière significative dans l intervale de température (moins de 18 C) etudiée. 20 th International Congress of Refrigeration, IIR / IIF, Sydney,