- Email: [email protected]
Food processing: The use of non-fouling food grade heat transfer fluids
Applied Thermal Engineering 84 (2015) 94e103
Contents lists available at ScienceDirect
Applied Thermal Engineering journal homepage: www.elsevier.com/locate/apthermeng
Research paper
Food processing: The use of non-fouling food grade heat transfer fluids Christopher Ian Wright a, *, Thomas Bembridge b, 1, Eole Picot c, 1, Julien Premel c, 1 a
Global Group, Cold Meece Estate, Cold Meece, Staffordshire ST15 0SP, United Kingdom Kinetic Partners, 1 London Wall, London EC2Y 5HB, UK c Pole Etudes Locomotive et TGV Optimisation de la Maintenance TGV chez SNCF, Paris, France b
h i g h l i g h t s Food grade heat transfer fluid (HTF) is colourless, non-toxic and non-irritating. This HTF is non-fouling and less carbon forms. Such HTFs can be safely used in food processing if they are HT-1 certified. A number of controls (e.g., HACCP) are used to ensure such fluids are safe. An additional check is to sample fluids to ensure food grade fluids are being used.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 30 October 2014 Accepted 8 March 2015 Available online 1 April 2015
It is reported that there are some 4000 companies operating high temperature thermal fluid systems in the UK and Ireland. This excludes steam or water based systems. The heat transfer fluids (HTFs) used in food processing are highly refined mineral HTFs that are non-toxic, non-irritating and lack an odour. If an HTF has been certified for use in food processing it carries an HT-1 certificate. HTFs suitable for use in food processing are commonly referred to as ‘non-fouling’ which means as they thermally degrade they produce small carbon particles that are suspended in the HTF. Moreover, the carbon formations are less sticky and this reduces the extent of adhesion to the internal surfaces of an HTF system. The current paper analysed the test reports from 1223 HTF systems and showed that, on average, the carbon residue for food grade HTF was lower than non-food grade HTF. This clearly demonstrates what the non-fouling nature of a food grade HTF. This paper then explored the regulatory, legal and environmental landscape for food grade HTFs. In this area of manufacturing, it is critical that the HTFs used are suitable for incidental contact with food. Other measures put consumer safety at the heart of all operations (i.e., internal company procedures such as hazard analysis and critical control points [HACCP]) and that food is safe for consumer consumption (e.g., external controls such as auditing manufacturers to ensure good quality and distribution practice). The authors introduce the idea that safety could be further enhanced through independent HTF sampling and chemical analysis of HTFs to ensure they are food grade and should be done without any interruption to a manufacturer's production. © 2015 Published by Elsevier Ltd.
Keywords: Heat transfer fluid Food grade Food processing Manufacturing
1. Introduction
Abbreviations: HTF, heat transfer fluid; TAN, total acid number; SOP, standard operating procedure; HACCP, hazard analysis and critical control point; REACH, registration, evaluation, authorisation and restriction of chemicals; CLP, classification, labelling and packaging; BRC, British retail consortium. * Corresponding author. Tel.: þ44 7967 230 155. E-mail address: [email protected] (C.I. Wright). 1 Previously employed by Global Group. http://dx.doi.org/10.1016/j.applthermaleng.2015.03.033 1359-4311/© 2015 Published by Elsevier Ltd.
A recent report estimated that the value of the heat transfer fluid (HTF) market in 2011 was $1684.0 million and by 2017 this will increase to $2557.2 million [1]. Heat transfer refers to the transfer of thermal energy. In this process, HTFs are a heat carrier between a heater and a heat consumer and back [2]. This is the basic requirement of a wide variety of industrial processes and the principle behind indirect heat transfer plants. HTFs are used in a
C.I. Wright et al. / Applied Thermal Engineering 84 (2015) 94e103
wide array of industries and in multiple applications. These are briefly outlined by industry in Table 1 [3] with the main focus being given to the food industry as this is the subject of this paper. Highly refined mineral HTFs are food grade and designed to be used in the processing of food as they are non-toxic, non-irritating and have no odour [4,5]. Food grade HTFs consist of a complex combination of hydrocarbons obtained from the intensive treatment of a petroleum fraction with sulphuric acid and oleum, by hydrogenation or by a combination of hydrogenation and acid treatment. A food grade HTF consists of saturated hydrocarbons having carbon numbers predominantly in the range C15eC50 [4]. Such HTFs are typically associated with fewer handling complaints such as occurs when they are spilled or leak, which is potentially one way that an HTF comes into contact with food. An important feature of food grade HTFs is that they are nonfouling, which means as they thermally degrade they produce small carbon particles that are suspended in the HTF and so carbon formations are less sticky and adhesions to the internal surfaces of an HTF system are reduced [6]. It is important to remember that all HTFs undergo thermal degradation over time. Indeed, at high operating temperatures, the bonds that exist between hydrocarbon chains will start to break and form shorter (commonly referred to as ‘light-ends’) and longer chained hydrocarbons (‘heavy-ends’) [7e9]. Both of these products have consequences for the safety of an HTF system and for HTF sampling, but the key point is that chemical analysis can be used to routinely assess the condition of HTFs. The build-up of light-end components is a potential fire risk as they decrease the ignition temperature of the HTF [9]. The accumulation of heavy-ends results in the formation of sticky carbon deposits or sludge and can be monitored by analysing the carbon residue in an HTF. The same is true for as oxidation of an HTF leads to the formation of carbon sludge and acids and these are therefore
also routinely part of the testing conducted on the condition of an HTF in the laboratory [8]. The formation of sludge or soft carbon will lead to a carbon coat forming on the internal surfaces of the HTF system and with time this will harden. The insulating effect of carbon means that the HTF system is less efficient (i.e., heat transfer rates are reduced resulting in longer heat-up time and lower production rates). Also, the hard carbon deposits work to form hot spots and have the potential to cause heater tubes and electrical elements in the HTF system to burn-out. The non-fouling nature of food grade HTFs indicates that the physical characteristics of HTFs differ. Therefore, the first aim of this research was to compare the non-fouling nature of a food grade HTF with that of a non-food grade HTF. This was done to assess whether differences in carbon levels exist between these fluids and what this looks like when tested in the laboratory. The second aim was outline the audits and control processes involved in the supply of food grade HTFs. Indeed, the supply of food grade HTFs is highly regulated in Europe and companies must comply with REACH and CLP regulations [6,10]. In addition, a manufacturer should source food grade HTFs that are suitable for use in food processing. Food grade HTFs carry an HT-1 certificate that is issued by governing bodies such as the NSF, which means the HTF contains ingredients considered safe for incidental contact with food [11,12]. Furthermore, a list of recommended food grade HTFs is commonly defined by insurers [13] and food retailers. Empirical evidence suggests that insurers normally define a list of suitable HTFs as opposed to recommending a single HTF [13]. Furthermore, manufacturers are audited to ensure that an appropriate HTF is being used in the processing of food. However, the current framework of self-governance and external audits means that there is a potential for companies to use HTFs not suitable for incidental contact with food. One justification being that a company may choose to dispose all foods should they come into contact
Table 1 Heat consumers by industry. Industry
Heating of (example)
Chemical and plastics Bitumen and tar processing Mineral oil Rubber
Distillation plants Bitumen tank storage Heavy oil plants in processing, storage, transport and in transfer stations Plants for the production of synthetic rubber
Food
1. Large kitchens 2. Pommes-frites plants 3. Potato chips plants 4. Plants for fat hardening and rendering 5. Bottle cleaning plants 6. Can washing plants 7. Spray drying plants for milk powder and blood powder production 8. Roller driers 9. Fat and oil vats 10. Baking ovens 11. Plants for the production of sweets 12. Plants for chocolate production 13. Plants for starch drying 14. Plants for the removal of smelly waste gases
Soaps and detergents Wood Paper Construction, stones and earths Textile Metal Electrical Shipping Aircraft and airlines Heating heat supply Energy generation
Spray and drying towers Press boards, plywood and veneer presses Coating rollers Ore preparation plants Tensioning frames for drying and/or fixing Acid and pickling tanks Drying cabinets Heavy oil and bitumen tanks on ships Galvanising tanks Warm water generators Concentrated solar plants
Extracted from Ref. [3].
95
96
C.I. Wright et al. / Applied Thermal Engineering 84 (2015) 94e103
Table 2 Chemical analysis routinely conducted in the laboratory. Variable analysed
Unit or element
Test method
Appearance Carbon residue Total acid number Strong acid number Closed flash point Open flash point Fire point Kinematic viscosity Water content Ferrous wear debris Elements (ppm < 5 microns)
Colour Percentage weight Milligram of potassium hydroxide per gram (mg KOH/g) Milligram of potassium hydroxide per gram (mg KOH/g) Temperature ( C) Temperature ( C) Temperature ( C) A square millimetre per second (i.e., mm2/s) Parts per million (ppm) Insolubles Iron, silicon
Coded according to colour IP14 IP139 IP177 ASTM D93 ASTM D92 ASTM D92 IP71 ASTM D6304 PQ Analex Method ASTM D5185
with an HTF. Thus meaning that an HTF approved for incidental contact would not be needed or even relevant to their business. The second aim of this paper was therefore to assess the internal and external controls involved in the supply of food grade HTFs to determine if routine sampling and chemical analysis of HTFs had a place in the current framework.
was not defined because a number of HTF systems were sampled and chemically analysed by other companies prior to Global Heat Transfer's involvement. This makes it difficult to assess the exact age of some HTFs. Furthermore, a number of interventions are used so sustain the condition of an HTF, for example percentage dilution, which impacts the reliability and validity of assessing age.
2. Methods
2.1. Sampling of heat transfer fluids
The source of data in this paper came from retrospective test reports gathered by Global Heat Transfer Ltd. between 2008 and 2013. Data was analysed if: i. A report contained a non-food grade HTF (specifically Globaltherm™ M) or a food grade HTF (specifically Globaltherm™ FG) [5,14]. These HTFs were chosen to ensure the HTFs being analysed were consistent in terms of chemical composition and physical properties, as well as being sourced from the same supply. Table 3 summarises the physical properties of the food grade and non-food grade HTF analysed in this paper. ii. There was at least one test report between 2008 and 2013. iii. Indirect heating systems were sampled. All test results were analysed and analysis did not factor the make of the HTF system or age of the HTF or HTF system. The age
All identified systems were sampled whilst the system was operating under normal operating conditions. A 500 ml sample of the HTF was withdrawn from the HTF system using a custom designed sampling device that isolated the sampled HTF so a true and representative sample of the HTF was obtained. This procedure is described in detail elsewhere [7,8]. 2.2. Chemical analysis of heat transfer fluids Table 2 summarises the test parameters and test methods routinely conducted in the laboratory by Global Heat Transfer. These tests were conducted according to ISO14001 and ISO17025 standards [15,16]. The key parameters measured were carbon residue, total acid number (TAN), closed flash temperature, open flash temperature, kinematic viscosity, water content, ferrous wear debris (insolubles) and elements (e.g., iron and silicon).
Table 3 Characteristics of virgin food grade and non-food grade mineral heat transfer fluids. Parameter
Unit
Globaltherm™ M*
Globaltherm™ FG**
Base liquid
Description
#Mineral oil base
Grade
Description
Non-food grade
Synthetic base,a hydro treated, highly refined mineral oil base Food grade
Other products with similar characteristics
Description
BP Transcal N Castrol Perfecto HT Gulftherm 32 Shell HT S2
Calflo Purity FG Multitherm PG1 Marlotherm FG Paratherm NF Therminol XP
Appearance
Description
Colourless transparent liquid with no odour
Operating range Closed flash point Open flash point Kinematic viscosity at 40 C Auto-ignition point Pour point Maximum film temperature Boiling point at 1013 mbar
Viscous clear-yellow liquid with a mild odour 10 to 320 210 230 29.8 350 12 330 365
C C C mm2/s C C C C
20 to 326 210 216 29.8 350 29 343 371
* and **, values extracted from Globaltherm™ M and Globaltherm™ FG product datasheets [5,14]. Note: values for carbon residue, total acid number, strong acid number, fire point, water content, ferrous wear and insoluble were not reported in the product datasheets. Values for virgin HTFs are expected to be: carbon reside, 0% weight; total acid number, <0.05 mg KOH/g; strong acid number, 0 mg KOH/g; fire point, ~255 C; water content, <100 ppm; ferrous wear, <10; insolubles less than 5 microns, 0 ppm. a Extracted from Ref. [3].
C.I. Wright et al. / Applied Thermal Engineering 84 (2015) 94e103
2.3. Data analysis Data was analysed using a number of statistical tests. A Pearson's correlation test was used to assess the linearity between two parameters [8]. The Pearson correlation coefficient (r) returns a score with þ1 indicating a positive correlation; 1 indicating a negative correlation; and, 0 indicating no correlation. The linearity was tested for significance to determine if a true correlation existed (P < 0.05) or not (P 0.05). Correlation coefficient values are reported along with 95% confidence intervals (CI), which were calculated using a 2-sided Fisher's Z test, as well as degrees of freedom (DF). A chi-squared test (c2) was used to assess the frequency and distribution of ratings (i.e., satisfactory, caution, action and serious) of parameters and to determine if categorised parameters were related or not. In some cases test parameters values were missing. Data is presented as mean ± standard deviation (SD) unless otherwise stated. All analyses and calculations were conducted using Microsoft Excel 2007 and/or Analyse-it® version 3.71. Statistical significance was demonstrated when the P-value was taken at the 5% level (i.e., P < 0.05). The number of post-hoc comparisons was factored into the linear correlation test to account for the multiple comparisons [17].
97
found between carbon residue and TAN. Further comparisons revealed a number of moderate (r ±0.25 and <±0.5; P < 0.05) and weak correlations (r < ±0.25; P < 0.05) that were not consistent for both HTFs. These correlations are detailed below. 3.1.2.1. Non-food grade heat transfer fluid. Moderate positive linear correlations were found when kinematic viscosity was compared with carbon residue and TAN. Weak negative linear correlations were found when elements were compared with carbon residue and TAN. Weak negative correlations were found when TAN was compared with fire point and open flash point. 3.1.2.2. Food grade heat transfer fluid. Weak negative correlations were found when TAN was compared with open and closed flash points. 3.1.3. The relationships between fluid type, carbon residue and total acid number as determined by frequency distribution Table 4 highlights that mean and peak values for carbon residue and TAN were higher for the non-food grade HTF. Chi-squared tests were therefore used to assess: i) the association between ratings of carbon residue and TAN by HTF type; ii) the association between ratings of carbon residue by HTF type; and, iii) the association between ratings of TAN and HTF type (see Tables 6 and 7).
3. Results A total of 1223 HTF samples were included in this analysis e 943 mineral HTF and 280 food grade HTF. These were taken from 97 customers (24 food grade HTF and 73 mineral HTF) with multiple HTF system lines based at multiple locations in the UK and Ireland. 3.1. Comparison of non-food grade and food grade heat transfer fluids 3.1.1. Group data obtained for non-food grade and food grade heat transfer fluids Table 4 compares the mean ± SD and ranges for all measured parameters. The greatest differences between these HTFs was observed for carbon residue (0.12% weight or 118.0%); TAN (0.05 mg KOH/g or 40.0%); closed flash point (16.9 C or 12.1%); water content (þ49.0 ppm or þ35.6%); and, elements (10.3 ppm less than 5 microns or 186.9%). These parameters all differed by more than 10% with water content being the only parameter that was higher for the food grade HTF. 3.1.2. Linear comparison of test parameters Table 5 presents the correlations between parameters. For both HTFs, strong positive linear correlations (r ±0.5; P < 0.05) were
3.1.3.1. The frequency distribution of ratings of carbon residue and TAN by fluid type. In keeping with the linear correlations reported above, a significant association was found between the ratings of carbon residue and TAN (P < 0.05; Table 6a and b). Indeed, for both HTFs, the overall rating (i.e., totals for rows and columns) for carbon residue and TAN decreased as the overall condition of the HTF worsened. Table 6a and b was analysed to determine if individual cells in these contingency tables were greater than the critical value (i.e., 22.46). For the non-food grade HTF there were three cells that exceeded the critical value: 143.1 (carbon residue rating, 0.75 to <1.0; TAN rating, 0.4 to <1.0); 43.3 (carbon residue rating, 1.0; TAN rating, 0.4 to <1.0); and, 31.4 (carbon residue rating, 0.5 to <0.75; TAN rating, 0.2 to <0.4). For the food grade HTF only one cell exceeded the critical value: 107.2 (carbon residue rating, 1.0; TAN rating, 0.4 to <1.0). Thus suggesting an association between higher ratings of carbon residue (0.75 for non-food HTF; 1.0 for food grade HTF) and TAN (0.4 to <1.0 for both HTFs). The last point to make is the percentage of HTF systems that had a satisfactory rating. Table 6a and b revealed that 67% (631 of 936) of non-food grade HTF systems had a satisfactory rating (i.e., carbon residue <0.5; TAN <0.2) compared with 81% (224 of 278) of food
Table 4 Analysis results obtained for non-food grade and food grade heat transfer fluids (HTFs) e grouped data. Variable analysed
Non-food grade HTF (n ¼ 943)
Food grade HTF (n ¼ 279)
Absolute difference, % (food grade minus non-food grade)
Carbon residue (% weight) Total acid number (mg KOH/g) Strong acid number (mg KOH/g) Closed flash point ( C) Open flash point ( C) Fire point ( C) Kinematic viscosity at 40 C (mm2/s) Water content (ppm) Ferrous wear debris Elements (ppm < 5 microns)
0.23 ± 0.25 (range, 0.01e2.94) 0.18 ± 0.16 (range, 0.01e2.34) 0.00 ± 0.00 (range, 0.00e0.00) 156.3 ± 33.1 (range, 46.0e268.0) 203.1 ± 18.7 (range, 20.0e254.0) 241.52 ± 14.83 (range, 84.0e271.0) 30.5 ± 4.1 (range, 3.1e85.6)
0.10 ± 0.19 (range, 0.00e1.76) 0.13 ± 0.09 (range, 0.01e0.55) 0.00 ± 0.00 (range, 0.00e0.00) 139.4 ± 34.8 (range, 34.0e211.0) 199.4 ± 18.8 (range, 129.0e235.0) 244.0 ± 22.8 (range, 180.0e547.0) 31.4 ± 2.6 (range, 18.5e43.0)
0.12, 118.0 0.05, 40.0 0.00, 0.00 16.9, 12.1 3.7, 1.9 0.6, 0.3 0.9, 2.9
88.7 ± 266.6 (range, 10.0e4200.0) 14.4 ± 18.6 (range, 8.0e217.0) 15.9 ± 28.9 (range, 0.0e307.0)
137.7 ± 711.9 (range, 22.0e10,000) 15.9 ± 39.9 (range, 0.0e632.0) 5.5 ± 12.3 (range, 0.0e93.0)
49.0, 35.6 1.4, 9.0 10.3, 186.9
Means ± SD.
98
C.I. Wright et al. / Applied Thermal Engineering 84 (2015) 94e103
Table 5 Linear comparison of test parameters recorded for a non-food grade and food grade heat transfer fluid (HTF). Parameter a. Non-food grade HTF Carbon
TAN
b. Food grade HTF Carbon
TAN
Comparison
Pearson correlation coefficient (r)
Fisher 95% CI
DF
P-value
vs. vs. vs. vs. vs. vs. vs. vs.
TAN closed flash point open flash point elements ferrous wear debris kinematic viscosity water content fire point
þ0.680 0.012 0.068 þ0.190 þ0.060 þ0.360 þ0.010 0.067
0.640 to 0.709 0.076 to 0.052 0.131 to 0.004 0.132 to 0.255 0.003 to 0.124 0.300 to 0.412 0.050 to 0.079 0.132 to 0.001
934 934 933 934 934 934 921 887
<0.0001y 0.709 0.038 <0.0001y 0.063 <0.0001y 0.663 0.047
vs. vs. vs. vs. vs. vs. vs.
fire point water content kinematic viscosity ferrous wear debris elements open flash point closed flash point
0.091 þ0.030 þ0.270 þ0.050 þ0.140 0.091 0.011
0.156 to 0.025 0.039 to 0.090 0.206 to 0.325 0.010 to 0.117 0.081 to 0.206 0.154 to 0.027 0.075 to 0.053
889 924 937 938 937 935 936
0.007z 0.435 <0.0001z 0.099 <0.0001z 0.005z 0.747
vs. vs. vs. vs. vs. vs. vs. vs.
TAN closed flash point open flash point elements ferrous wear debris kinematic viscosity water content fire point
þ0.510 þ0.000 0.091 þ0.160 þ0.060 þ0.160 þ0.050 0.062
0.416 to 0.591 0.114 to 0.121 0.206 to 0.028 0.40 to 0.270 0.060 to 0.175 0.043 to 0.272 0.074 to 0.162 0.181 to 0.059
276 277 275 276 277 277 274 263
<0.0001y 0.950 0.133 0.009 0.332 0.007 0.459 0.319
vs. vs. vs. vs. vs. vs. vs.
fire point water content kinematic viscosity ferrous wear debris elements open flash point closed flash point
0.109 þ0.040 0.041 þ0.110 þ0.010 0.179 0.161
0.227 0.076 0.158 0.008 0.109 0.291 0.274
262 273 276 276 275 274 276
0.077 0.484 0.491 0.069 0.884 0.003z 0.007z
to to to to to to to
0.012 0.160 0.077 0.224 0.127 0.062 0.044
y (n ¼ 8) and z (n ¼ 7) comparisons were significant even when the number of post-hoc comparisons was factored into the analysis (i.e., P < 0.05 divided by the number of posthoc comparisons) [17]. TAN, total acid number. DF, degrees of freedom.
grade HTF systems. Further analysis revealed that 69% (648 of 936) and 81% (225 of 278) of non-food grade and food grade HTF systems had satisfactory ratings for TAN. These satisfactory ratings were even higher for carbon residue where satisfactory ratings were 80% (840 of 936) and 98% (272 of 278) for non-food grade and food grade HTFs respectively. 3.1.3.2. The relationship between carbon residue and fluid type. The chi-squared test showed a significant relation between carbon residue and HTF type (P < 0.05; Table 7a). Furthermore, the contingency table showed the frequency of satisfactory ratings was highest and a serious rating was the lowest. For the non-food grade HTF, the number of satisfactory ratings was lower than expected (i.e., observed minus expected) and higher than expected for caution, action and serious ratings. This was completely the opposite for the food grade HTF. 3.1.3.3. The relationship between total acid number and fluid type. This analysis revealed the same trends for non-food grade and food grade HTFs as was found for carbon residue (please see Table 7b). 3.2. The internal and external controls involved in the supply of food grade 3.2.1. Audits and controls involved in the supply of food grade HTFs A number of regulations e such as REACH [6] and CLP [10] e ensure that chemicals are distributed with sufficient warnings so they are safely managed and used (see Table 8). In the context of
food processing, HACCP [18] and quality/distribution audits [19] are used to ensure that food is managed safely both during processing and then when being distributed to the consumer. An additional safety measure is provided by trade bodies such as the British Retail Consortium [20]. This trade association is involved in all types of retail including food retailers (e.g., supermarkets) and promote responsible retailing. Furthermore, industrial insurers, such as FM Global [13], work closely with manufacturers to make sure that commercial operations are adequately insured. The manufacturer is central to all the activities defined in Fig. 2 yet there is no mechanism to ensure that food grade HTFs are used in the processing of food. Even though at this point in the process there is the potential for the food to come into contact with the HTF. The HACCP is an internal document used to define the acceptable contamination limits for a food, but the manufacturer may choose to dispose of all food that comes into contact with an HTF. For this reason a manufacturer may choose to use a non-food grade HTF as opposed to a food grade HTF. In another scenario, the retailer may stipulate the use of a food grade HTF and confirm its use through audits of the manufacturing facility. However, the HTF would not be sampled and chemically analysed and confirmation would be based on paperwork. Lastly, the insurer could stipulate the use of a food grade HTF, but, again, confirmation that a food grade HTF is used would be based on a paper-based audit. Insurers also assess the frequency that an HTF is sampled and chemically analysed to ensure the flash point temperature of the HTF is being managed as stipulated [7e9]. However, at no point would the insurer call request the HTF to be sampled and chemically analysed to validate documentation.
C.I. Wright et al. / Applied Thermal Engineering 84 (2015) 94e103
99
Table 6 The frequency distribution of carbon residue and total acid number (TAN) for non-food grade and food grade heat transfer fluids (HTFs). Calculation
TAN rating
Carbon residue rating
Totals
<0.5
0.5 to <0.75
0.75 to <1.0
1.0
<0.2
631.0 581.5 49.5 2446.4 4.2
12.0 30.5 18.5 340.8 11.2
2.0 20.8 18.8 352.3 17.0
3.0 15.2 12.2 149.6 9.8
648.0
Observed (O) Expected (E) OE O E2 (O E2)/E
0.2 to <0.4
174.0 193.8 19.8 393.9 2.0
28.0 10.2 17.8 318.5 31.4
8.0 6.9 1.1 1.2 0.2
6.0 5.1 0.9 0.9 0.2
216.0
Observed (O) Expected (E) OE O E2 (O E2)/E
0.4 to <1.0
35.0 61.9 26.9 724.9 11.7
4.0 3.2 0.8 0.6 0.2
20.0 2.2 17.8 316.4 143.1
10.0 1.6 8.4 70.2 43.3
69.0
Observed (O) Expected (E) OE O E2 (O E2)/E
1.0
0.0 2.7 2.7 7.2 0.0
0.0 0.1 0.1 0.0 0.0
0.0 0.1 0.1 0.0 0.0
3.0 0.1 2.9 8.6 0.0
3.0
840.0
44.0
30.0
22.0
936.0 274.2 9 22.46 <0.001
<0.2
224.0 220.1 3.9 14.9 0.1
1.0 2.4 1.4 2.0 0.8
0.0 0.8 0.8 0.7 0.8
0.0 1.6 1.6 2.6 1.6
225.0
Observed (O) Expected (E) OE O E2 (O E2)/E
0.2 to <0.4
45.0 47.0 2.0 3.9 0.1
2.0 0.5 1.5 2.2 4.2
1.0 0.2 0.8 0.7 4.0
0.0 0.3 0.3 0.1 0.3
48.0
Observed (O) Expected (E) OE O E2 (O E2)/E
0.4 to <1.0
3.0 4.9 1.9 3.6 0.7
0.0 0.1 0.1 0.0 0.1
0.0 0.0 0.0 0.0 0.0
2.0 0.0 2.0 3.9 107.2
5.0
Observed (O) Expected (E) OE O E2 (O E2)/E
1.0
0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0
0.0
272.0
3.0
1.0
2.0
278.0 120.0 9 22.46 <0.001
a. Non-food grade HTF Observed (O) Expected (E) OE O E2 (O E2)/E
Totals c2 value Degrees of freedom Critical value P-value b. Food grade HTF Observed (O) Expected (E) OE O E2 (O E2)/E
Totals c2 value Degrees of freedom Critical value P-value
Condition ratings for carbon residue: satisfactory, <0.5; caution, 0.5 to <0.75; action, 0.75 to <1.0; and, serious, 1.0. Condition ratings for TAN: satisfactory, <0.2; caution, 0.2 to <0.4; action, 0.4 to <1.0; and, serious, 1.0. Grey shaded area represents concurrent condition ratings for carbon residue and TAN.
4. Discussion Mean data clearly showed mean and maximum values for carbon residue and TAN were higher for non-food grade HTF than for the food grade HTF. Individual data was then explored and linear analysis showed there was a strong positive relationship between carbon residue and TAN for both HTFs. Subsequently, frequency distribution analysis, using chi-squared tests, suggested a possible association between high ratings of carbon residue (0.75 for non-
food HTF; 1.0 for food grade HTF) and relatively moderate ratings of TAN (0.4 to <1.0 for both HTFs). Meaning that carbon residue tended to be higher relative to TAN. The effect of HTF type was explored and showed that satisfactory ratings were highest and serious ratings were lowest and this was true for both HTFs. The number of satisfactory ratings was higher than expected for the food grade HTF and lower than expected for the non-food grade HTF. Thus meaning that non-food grade HTFs had a relatively higher number of scores in other categories judged to be not
100
C.I. Wright et al. / Applied Thermal Engineering 84 (2015) 94e103
Table 7 Determination of the relationship between fluid type and carbon residue and total acid number. HTF
Calculation
Carbon residue rating <0.5
0.75 to <1.0
1.0
45.0 37.0 8.0 64.1 1.7
30.0 23.9 6.1 37.3 1.6
23.0 20.0 3.0 8.8 0.4
938.0
272.0 254.9 17.1 291.4 1.1
3.0 11.0 8.0 64.1 5.8
1.0 7.1 6.1 37.3 5.2
3.0 6.0 3.0 8.8 1.5
279.0
1112.0
48.0
31.0
26.0
1217.0 15.8 3 12.84 <0.005
a. The relationship between carbon residue rating and fluid type Non-food grade Observed (O) 840.0 Expected (E) 857.1 OE 17.1 2 OE 291.4 (O E2)/E 0.3 Food grade
Observed (O) Expected (E) OE O E2 (O E2)/E
Totals c2 value Degrees of freedom Critical value P-value HTF
Calculation
TAN rating <0.2
Observed (O) Expected (E) OE O E2 (O E2)/E
Totals c2 value Degrees of freedom Critical value P-value
Totals 0.2 to <0.4
0.4 to <1.0
1.0
222.0 209.1 12.9 165.2 0.8
69.0 57.1 11.9 141.4 2.5
3.0 2.3 0.7 0.5 0.2
940.0
224.0 198.6 25.4 646.6 3.3
49.0 61.9 12.9 165.2 2.7
5.0 16.9 11.9 141.4 8.4
0.0 0.7 0.7 0.5 0.7
278.0
870.0
271.0
74.0
3.0
1218.0 19.4 3 16.27 <0.001
b. The relationship between total acid number (TAN) rating and fluid type Non-food grade Observed (O) 646.0 Expected (E) 671.4 OE 25.4 O E2 646.6 (O E2)/E 1.0 Food grade
Total 0.5 to <0.75
n, number of samples. NA, not applicable. Note: that in a small number of cases, values were missing and this explains why the totals here vary from those presented in the previous table. Condition ratings for carbon residue: satisfactory, <0.5; caution, 0.5 to <0.75; action, 0.75 to <1.0; and, serious, 1.0. Condition ratings for TAN: satisfactory, <0.2; caution, 0.2 to <0.4; action, 0.4 to<1.0; and, serious, 1.0.
Table 8 The intention and focus of various controls in the processing of food. Environment
Controls
Intention
Focus
Internal
SOP
Consistency of services and/or products
Internal
HACCP
External
REACH regulation
External
CLP legislation
External
Insurer
External
Quality and distribution audits
External
Consumer agencies
The documented processes that a company has in place and uses on a day-to-day basis to ensure services and/or products are delivered consistently across the organisation A system that helps food business operators look at how they handle food and introduces procedures to make sure the food produced is safe to eat Companies must identify and manage the risks linked to the substances they manufacture and market in the European Union. It applies to all chemical substances from those used in industrial processes to everyday fast-moving-consumer-goods. Companies need to demonstrate how to safely use the substances and this must be communicated to the users Ensures that the substances and mixtures placed on the market, across the globe, are classified, labelled and packaged appropriately Commercial and industrial insurers are used to manage risk needs (i.e., identification, assessment, avoidance and reduction) and protect the value created by the customer's organisation Examples include attainment and maintenance of ISO certification such as ‘ISO22000:2005 e Food Safety Management Systems’ which is a standard outlined for the control of food safety For example the BRC which is the UK trade association for the entire retail industry and includes food products sold to consumers
Managing food safely
Identification and management of risks associated with substances
Correct classification, labelling and packaging Managing risk
Managing food safely
Consumer safety
SOP, standard operating procedure; HACCP, hazard analysis and critical control point; REACH, registration, evaluation, authorisation and restriction of chemicals; CLP, classification, labelling and packaging; and, BRC, British retail consortium.
C.I. Wright et al. / Applied Thermal Engineering 84 (2015) 94e103
101
satisfactory. Finally, satisfactory ratings for carbon residue and TAN were markedly higher for food grade HTFs: carbon residue, 90% versus 98% (non-food versus food grade HTFs respectively); and, TAN, 69% versus 81%. 4.1. Comparison of non-food grade and food grade heat transfer fluids It is estimated that there are around 4000 companies using high temperature HTF systems in the UK and Ireland [21] with roughly one-in-five systems, based on the HTF systems that the authors is aware of, using HTFs to process food. Food grade HTFs are used in food processing as they are generally regarded as being safe for incidental contact. Another feature of food grade HTFs is that they are non-fouling and the analysis of this characteristic was the primary goal of this paper. Non-fouling is used to describe the carbon formatting capacity of an HTF. If carbon residue levels are reduced this means that there is less carbon present which means that less carbon adheres to the internal surfaces of an HTF system. The authors were unaware of comparisons between non-food grade and food grade HTFs so these HTFs were compared to assess mean carbon residue levels and ranges, and how these were distributed (i.e., if they are satisfactory or not). The first aspect of this analysis showed that carbon level was 0.10 and 0.23 percentage weight for food grade and non-food grade HTF. This reflects a percentage difference of 118% (0.12/0.10 * 100). TAN levels were also lower for food grade HTF (0.13 versus 0.18; food grade versus non-food grade HTF). Other differences between HTFs were identified (see Table 4). These included closed flash point, water content and elements, the latter two reflecting contamination and not a physical difference in the HTFs. The lower carbon residue level for food grade HTFs is very interesting because during thermal degradation of an HTF two products are formed e long (commonly referred to as ‘heavy ends’) and short chain (‘light ends’) hydrocarbons [9]. Testing is used to identify these components and usually seen as increased carbon residue and decreased closed flash point, which decreases as the composition of light ends in the HTF increases. With this in mind, it is interesting that non-food grade HTF had a higher carbon residue level but this was accompanied by a currently lower closed flash point as might have been expected. The second aspect of the analysis was to explore the relationship between the test parameters (see Table 5). This was done by constructing xey plots and assessing the correlation between parameters. The principle finding was the strong positive (r ±0.5; P < 0.05) correlation between carbon residue and TAN. This was seen for both HTFs. Other correlations revealed weaker correlations (r < ±0.5; P < 0.05). Interestingly, closed and open flash points were found to negatively relate to TAN for food grade HTF. For non-food grade HTF TAN was found to negatively relate to fire point and open flash point, but not closed flash point. This data demonstrates the existence of subtle differences between HTFs but also shows that the one consistent finding between HTFs was the existence of a relationship between carbon residue and TAN. This therefore formed the basis of the current comparison between non-food and food grade HTFs. Fig. 1 compares the plots for non-food grade and food grade HTFs. The equation for inserted lines are y ¼ 0.4435x þ 0.075 for non-food grade and y ¼ 0.2809x þ 0.0984 for the food grade HTF. So the question here is what is the relevance of this data? The equations suggest that for any given level of carbon residue, the TAN is lower for food grade HTFs. Conversely, this also means that for any given level of TAN, the carbon residue level is higher for non-food grade HTF. To understand how test report scores were distributed, carbon residue and TAN were categorised into four groups. This categorisation had a maximal score of 1 and reflected a serious level for
Fig. 1. Linear comparisons of carbon residue and total acid number (TAN) for a) nonfood grade and b) food grade heat transfer fluids (HTFs).
both parameters. Whereas a score close to zero reflected that seen for a virgin HTF. These categorisations are reported in Tables 6 and 7. This specific analysis was done to understand if carbon residue and TAN varied by HTF type and if this rating explained the lower mean carbon residue and TAN scores for food grade HTF. Initially HTFs were analysed separately so that carbon residue and TAN
Fig. 2. An illustration of the audits (internal and external) and controls involved in the food processing supply chain. SOP, standard operating procedure; HACCP, hazard analysis and critical control point; REACH, registration, evaluation, authorisation and restriction of chemicals; CLP, classification, labelling and packaging; and, BRC, British retail consortium.
102
C.I. Wright et al. / Applied Thermal Engineering 84 (2015) 94e103
could be compared by HTF type, and then together to see if carbon residue and TAN differed by HTF type. This analysis revealed an association between irrespective of how the data was analysed. Indeed, both approaches used chi-squared tests and showed that for both HTFs the number of satisfactory ratings (i.e., <0.5 for carbon residue and <0.2 for TAN) was higher than expected. Furthermore, the number of unsatisfactory ratings (i.e., any category that was not judged to be satisfactory) decreased as the condition rating progressive worsened. Thus meaning that the smallest tally of test results existed in the category where carbon residue and TAN was 1.0. Analysis also suggests a possible association between high ratings of carbon residue (0.75 for non-food HTF; 1.0 for food grade HTF) and relatively moderate ratings of TAN (0.4 to <1.0 for both HTFs). Meaning that carbon residue tended to be higher relative to TAN, which is also seen in the linear correlation analysis (see Fig. 1). Chi-squared testing was also used to assess the association between carbon residue and both HTFs (see Table 7). This showed that the number of test reports that had a satisfactory rating was higher than expected in the food grade HTF group and lower than expected in the non-food grade HTF group. Furthermore, the tallies of test reports with an unsatisfactory rating were higher for the non-food grade HTF and lower for the food grade HTF. The same finding was also true for assessments of TAN. This clearly shows that food grade HTF had more satisfactory ratings and fewer serious ratings than the non-food grade HTF. 4.2. The internal and external controls involved in the supply of food grade The second aim was to assess the landscape for audits and controls encompassed in the supply of food grade HTFs. The supply of food grade HTFs is highly regulated in Europe and companies must comply with REACH and CLP regulations [6,10]. These regulations also concern the safe distribution and handling of HTFs [6,10]. Other elements include the processing controls and safeguards such as HACCP [18] and quality/distribution audits [19]. The British Retail Consortium is a trade association [20] that is involved in the protection of the consumer and promotion of responsible retailing. Lastly, industrial insurers [13] that work to ensure HTFs and HTF systems are managed safely. In addition, a manufacturer should only source food grade HTFs that are suitable for food processing and incidental contact with food [17]. HT-1 certification is governed by two well-known bodies e the NSF and InS. In the case of the NSF, the components comprising a fluid are compared against the NSF's registration guidelines for proprietary substances and non-food compounds and 21 CFR [18]. Based on a toxicologists' analysis, an HTF is then registered as HT-1 [11], which is specific to food grade HTFs. In some cases, but not all, the use of food grade HTFs is stipulated by insurers and food retailers, and certain manufacturers will be routinely audited to ensure that an appropriate HTF is being used in the processing of food. However, the current framework means there is potential for companies to use HTFs not suitable for food processing. One justification being that a company may choose to dispose all foods if they come into contact with the HTF they are using. Thus meaning that an HTF approved for incidental contact would not be needed or even relevant to their business. Another example for not using food grade HTFs is that the HTF will not come into contact with food, which seems logical but caution should be taken as leaks can happen and go undetected until the consumer comes into contact with the processed food. Indeed, in 1998 more than 222,260 kg of smoked boneless hams were recalled by Smithfield Foods because they were contaminated by a gear lubricant after several customers reported a ‘bad taste’ and ‘burning in their throat’ which lasted anywhere up to 3 h [22]. The key take
home message being that an inappropriate HTF was used and that these adverse events could have been minimised if a food grade HTF had been used. A review of all processes, audits and controls indicates that there is a gap in current practice and this means that non-food grade HTFs may be used or that the HTF is never sampled and chemically analysed to confirm it is a food grade HTF. Such testing is quick and easy to conduct [7] and can be done without interrupting production. Indeed, it is possible for HTF systems to be sampled prior to and following any intervention that affects the HTF [8] and whilst the HTF system is running under normal conditions. Such testing should be followed-up with routine and regular HTF sampling to assess the effectiveness of any intervention and also to assess the ongoing state of the HTF [23]. A number of interventions can be used to help maintain the condition of the HTF such as the installation of a light-ends removal kit (LERK) which can be used be fitted permanently to the HTF system or used intermittently [9] in a similar manner to batch venting but without the need to interrupt system production [24]. Testing can also be used to assess the effectiveness of such interventions and to propose subsequent actions. It would seem logical that HTF sampling is performed by a specialist in this area to deter malpractice. The test results from such testing could then be provided to: 1) the insurer to show the HTF is being managed and the HTF system is safe; 2) retailers to demonstrate that an appropriate food grade HTF is being used in food processing; and, 3) external agencies to show that consumer safety is not being compromised. 5. Conclusions The primary aim of this paper was to compare the non-fouling nature of non-food grade and food grade HTFs. The take home message from this analysis is that non-fouling is observed as lower carbon residue and TAN levels. Indeed, mean carbon residue and TAN levels were higher for non-food grade HTF, but irrespective of this, carbon residue and TAN were found to be strongly related for both HTFs. Linear plots demonstrate that carbon residue was higher relative to TAN. Analysis looking at the frequency of distribution shows that ratings for HTFs were highest (i.e., observed tallies were higher than expected) in the category judged to be satisfactory. The same analysis technique also showed that food grade HTFs tended to have more scores in the satisfactory category rating than nonfood grade HTFs. Conversely, non-food grade HTF had more unsatisfactory ratings whereas the tallies for food grade HTF were the complete opposite (i.e., lower than expected). In fact, comparisons showed that in 69% and 81% of cases TAN was <0.2 (for non-food grade and food grade HTF, respectively). This difference is þ12% in favour of the food grade HTF. In 90% and 98% of cases, carbon residue was <0.5 e a difference of þ8% in favour of food grade HTF. The secondary aim was to analyse the legal, environmental and regulatory landscape concerning the use of food grade HTF in food processing. This revealed that there is no body physically checking that a food grade HTF is being used in the HTF system. One easy solution to consider is the random and independent sampling and chemical analysis of HTFs to ensure that a food grade HTF is being used and this would also work as an additional measure underpinning consumer safety. Acknowledgements The author would like to acknowledge the Global Heat Transfer team. Specifically the engineering team (Danny Bradford, Dave Dyer and Martyn Tinsley) and the technical support from Andy Burns, Lisa Cho and Ian Halliwell.
C.I. Wright et al. / Applied Thermal Engineering 84 (2015) 94e103
References [1] ReportsnReports.com Updates its Store with “Global Heat Transfer Fluid Market [Thermic Fluid/High Temperature/Synthetic Heat Transfer Fluids] By Product Type [Mineral Oils, Silicone & Aromatics, PAG & Glycol Based Products & Others], Applications & Geography e Forecasts To 2017” Market Research Report, Source: www.marketsandmarkets.com (accessed 22.10.14). [2] D.J. Kukulka, M. Devgun, Fluid temperature and velocity effect on fouling, Appl. Therm. Eng. 27 (2007) 2732e2744. [3] Wagner O. Walter, Heat transfer technique with organic media, in: Heat Transfer Media, second ed., Maria-Eich-Strabe, Graefelfing, Germany, 1997, pp. 4e58 (Chapter 2). [4] White Mineral Oil (petroleum), Source: http://apps.echa.europa.eu/registered/ data/dossiers/DISS-9ea08dfc-55a5-3699-e044-00144f67d031/DISS-9ea08dfc55a5-3699-e044-00144f67d031_DISS-9ea08dfc-55a5-3699-e04400144f67d031.html (accessed 22.10.14). [5] Globaltherm™ FG, Source: Globalheattransfer.co.uk (accessed 30.10.14). [6] REACH regulation, Source: http://echa.europa.eu/regulations/reach/ understanding-reach (accessed 21.10.14). [7] C.I. Wright, A. Burns, C.J. Jones, Sampling hot heat transfer fluids: simple insights for gaining a representative sample, Int. J. Eng. Innov. Technol. 3 (2013) 202e204. [8] C.I. Wright, Thermal heat transfer fluid problems following a system flush with caustic and water, Case Stud. Therm. Eng. 2 (2014) 91e94. [9] C.I. Wright, Effective management of heat transfer fluid flash point temperatures using a light-ends removal kit (LERK), Case Stud. Therm. Eng. 4 (2014) 9e14. [10] CLP legislation, http://echa.europa.eu/regulations/clp/legislation (accessed 21.10.14). [11] P. Bird, Advice on how to convert your processing plant to H1 and HT-1 registered food machinery lubricants, Food Market. Technol. (June 2009). Source: http://d-nb.info/994769512/34 (accessed 22.10.14).
103
[12] CFR e Code of Federal Regulations Title 21, Source: www.accessdata.fda.gov/ scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr¼178.3570 (accessed 22.10.14). [13] FM Global, Source: www.fmglobal.com/page.aspx?id¼01010000 (accessed 21.10.14). [14] Globaltherm™ M, Source: Globalheattransfer.co.uk (accessed 30.10.14). [15] ISO 14001, Environmental Management, 2004. Source: http://www.iso.org/ iso/iso14000 (accessed 22.10.14). [16] ISO/IEC 17025, General Requirements for the Competence of Testing and Calibration Laboratories, 2005. Source: http://www.standards.org/standards/ listing/iso_17025 (accessed 22.10.14). [17] C.I. Wright, H.J. Scholten, J.C. Schilder, B.M. Elsen, W. Hanselaar, C.I. Kroner, R. Draijer, J.J. Kastelein, W. Stok, J. Karemaker, E. de Groot, Arterial stiffness, endothelial function and microcirculatory reactivity in healthy young males, Clin. Physiol. Funct. Imaging 28 (5) (2008) 299e306. [18] Hazard Analysis and Critical Control Point (HACCP), www.food.gov.uk/ business-industry/caterers/haccp (accessed 21.10.14). [19] ISO 22000 e Food safety management, Source: http://www.iso.org/iso/home/ standards/management-standards/iso22000.htm (accessed 21.10.14). [20] British Retail Consortium (BRC), Source: www.brc.org.uk/brc_home.asp (accessed 21.10.14). [21] A. McKay, R. Franklin, Good practice: fire and explosion hazards with thermal fluid systems, Loss Prev. Bull. 219 (2011) 13e17. [22] S. Gebarin, The basics of food-grade lubricants, Machin. Lubric. (2009). Source: www.machinerylubrication.com/Read/1857/food-grade-lubricants-basics (accessed 22.10.14). [23] C.I. Wright, E. Picot, T. Bembridge, The relationship between the condition of a mineral-based heat transfer fluid and the frequency that it is sampled and chemically analysed, Appl. Therm. Eng. 75 (2015) 918e922. [24] C.I. Wright, J. Premel, Heat transfer system safety: comparing the effectiveness of batch venting and a light-ends removal kit (LERK), Case Stud. Therm. Eng. 4 (2014) 215e221.
Contents lists available at ScienceDirect
Applied Thermal Engineering journal homepage: www.elsevier.com/locate/apthermeng
Research paper
Food processing: The use of non-fouling food grade heat transfer fluids Christopher Ian Wright a, *, Thomas Bembridge b, 1, Eole Picot c, 1, Julien Premel c, 1 a
Global Group, Cold Meece Estate, Cold Meece, Staffordshire ST15 0SP, United Kingdom Kinetic Partners, 1 London Wall, London EC2Y 5HB, UK c Pole Etudes Locomotive et TGV Optimisation de la Maintenance TGV chez SNCF, Paris, France b
h i g h l i g h t s Food grade heat transfer fluid (HTF) is colourless, non-toxic and non-irritating. This HTF is non-fouling and less carbon forms. Such HTFs can be safely used in food processing if they are HT-1 certified. A number of controls (e.g., HACCP) are used to ensure such fluids are safe. An additional check is to sample fluids to ensure food grade fluids are being used.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 30 October 2014 Accepted 8 March 2015 Available online 1 April 2015
It is reported that there are some 4000 companies operating high temperature thermal fluid systems in the UK and Ireland. This excludes steam or water based systems. The heat transfer fluids (HTFs) used in food processing are highly refined mineral HTFs that are non-toxic, non-irritating and lack an odour. If an HTF has been certified for use in food processing it carries an HT-1 certificate. HTFs suitable for use in food processing are commonly referred to as ‘non-fouling’ which means as they thermally degrade they produce small carbon particles that are suspended in the HTF. Moreover, the carbon formations are less sticky and this reduces the extent of adhesion to the internal surfaces of an HTF system. The current paper analysed the test reports from 1223 HTF systems and showed that, on average, the carbon residue for food grade HTF was lower than non-food grade HTF. This clearly demonstrates what the non-fouling nature of a food grade HTF. This paper then explored the regulatory, legal and environmental landscape for food grade HTFs. In this area of manufacturing, it is critical that the HTFs used are suitable for incidental contact with food. Other measures put consumer safety at the heart of all operations (i.e., internal company procedures such as hazard analysis and critical control points [HACCP]) and that food is safe for consumer consumption (e.g., external controls such as auditing manufacturers to ensure good quality and distribution practice). The authors introduce the idea that safety could be further enhanced through independent HTF sampling and chemical analysis of HTFs to ensure they are food grade and should be done without any interruption to a manufacturer's production. © 2015 Published by Elsevier Ltd.
Keywords: Heat transfer fluid Food grade Food processing Manufacturing
1. Introduction
Abbreviations: HTF, heat transfer fluid; TAN, total acid number; SOP, standard operating procedure; HACCP, hazard analysis and critical control point; REACH, registration, evaluation, authorisation and restriction of chemicals; CLP, classification, labelling and packaging; BRC, British retail consortium. * Corresponding author. Tel.: þ44 7967 230 155. E-mail address: [email protected] (C.I. Wright). 1 Previously employed by Global Group. http://dx.doi.org/10.1016/j.applthermaleng.2015.03.033 1359-4311/© 2015 Published by Elsevier Ltd.
A recent report estimated that the value of the heat transfer fluid (HTF) market in 2011 was $1684.0 million and by 2017 this will increase to $2557.2 million [1]. Heat transfer refers to the transfer of thermal energy. In this process, HTFs are a heat carrier between a heater and a heat consumer and back [2]. This is the basic requirement of a wide variety of industrial processes and the principle behind indirect heat transfer plants. HTFs are used in a
C.I. Wright et al. / Applied Thermal Engineering 84 (2015) 94e103
wide array of industries and in multiple applications. These are briefly outlined by industry in Table 1 [3] with the main focus being given to the food industry as this is the subject of this paper. Highly refined mineral HTFs are food grade and designed to be used in the processing of food as they are non-toxic, non-irritating and have no odour [4,5]. Food grade HTFs consist of a complex combination of hydrocarbons obtained from the intensive treatment of a petroleum fraction with sulphuric acid and oleum, by hydrogenation or by a combination of hydrogenation and acid treatment. A food grade HTF consists of saturated hydrocarbons having carbon numbers predominantly in the range C15eC50 [4]. Such HTFs are typically associated with fewer handling complaints such as occurs when they are spilled or leak, which is potentially one way that an HTF comes into contact with food. An important feature of food grade HTFs is that they are nonfouling, which means as they thermally degrade they produce small carbon particles that are suspended in the HTF and so carbon formations are less sticky and adhesions to the internal surfaces of an HTF system are reduced [6]. It is important to remember that all HTFs undergo thermal degradation over time. Indeed, at high operating temperatures, the bonds that exist between hydrocarbon chains will start to break and form shorter (commonly referred to as ‘light-ends’) and longer chained hydrocarbons (‘heavy-ends’) [7e9]. Both of these products have consequences for the safety of an HTF system and for HTF sampling, but the key point is that chemical analysis can be used to routinely assess the condition of HTFs. The build-up of light-end components is a potential fire risk as they decrease the ignition temperature of the HTF [9]. The accumulation of heavy-ends results in the formation of sticky carbon deposits or sludge and can be monitored by analysing the carbon residue in an HTF. The same is true for as oxidation of an HTF leads to the formation of carbon sludge and acids and these are therefore
also routinely part of the testing conducted on the condition of an HTF in the laboratory [8]. The formation of sludge or soft carbon will lead to a carbon coat forming on the internal surfaces of the HTF system and with time this will harden. The insulating effect of carbon means that the HTF system is less efficient (i.e., heat transfer rates are reduced resulting in longer heat-up time and lower production rates). Also, the hard carbon deposits work to form hot spots and have the potential to cause heater tubes and electrical elements in the HTF system to burn-out. The non-fouling nature of food grade HTFs indicates that the physical characteristics of HTFs differ. Therefore, the first aim of this research was to compare the non-fouling nature of a food grade HTF with that of a non-food grade HTF. This was done to assess whether differences in carbon levels exist between these fluids and what this looks like when tested in the laboratory. The second aim was outline the audits and control processes involved in the supply of food grade HTFs. Indeed, the supply of food grade HTFs is highly regulated in Europe and companies must comply with REACH and CLP regulations [6,10]. In addition, a manufacturer should source food grade HTFs that are suitable for use in food processing. Food grade HTFs carry an HT-1 certificate that is issued by governing bodies such as the NSF, which means the HTF contains ingredients considered safe for incidental contact with food [11,12]. Furthermore, a list of recommended food grade HTFs is commonly defined by insurers [13] and food retailers. Empirical evidence suggests that insurers normally define a list of suitable HTFs as opposed to recommending a single HTF [13]. Furthermore, manufacturers are audited to ensure that an appropriate HTF is being used in the processing of food. However, the current framework of self-governance and external audits means that there is a potential for companies to use HTFs not suitable for incidental contact with food. One justification being that a company may choose to dispose all foods should they come into contact
Table 1 Heat consumers by industry. Industry
Heating of (example)
Chemical and plastics Bitumen and tar processing Mineral oil Rubber
Distillation plants Bitumen tank storage Heavy oil plants in processing, storage, transport and in transfer stations Plants for the production of synthetic rubber
Food
1. Large kitchens 2. Pommes-frites plants 3. Potato chips plants 4. Plants for fat hardening and rendering 5. Bottle cleaning plants 6. Can washing plants 7. Spray drying plants for milk powder and blood powder production 8. Roller driers 9. Fat and oil vats 10. Baking ovens 11. Plants for the production of sweets 12. Plants for chocolate production 13. Plants for starch drying 14. Plants for the removal of smelly waste gases
Soaps and detergents Wood Paper Construction, stones and earths Textile Metal Electrical Shipping Aircraft and airlines Heating heat supply Energy generation
Spray and drying towers Press boards, plywood and veneer presses Coating rollers Ore preparation plants Tensioning frames for drying and/or fixing Acid and pickling tanks Drying cabinets Heavy oil and bitumen tanks on ships Galvanising tanks Warm water generators Concentrated solar plants
Extracted from Ref. [3].
95
96
C.I. Wright et al. / Applied Thermal Engineering 84 (2015) 94e103
Table 2 Chemical analysis routinely conducted in the laboratory. Variable analysed
Unit or element
Test method
Appearance Carbon residue Total acid number Strong acid number Closed flash point Open flash point Fire point Kinematic viscosity Water content Ferrous wear debris Elements (ppm < 5 microns)
Colour Percentage weight Milligram of potassium hydroxide per gram (mg KOH/g) Milligram of potassium hydroxide per gram (mg KOH/g) Temperature ( C) Temperature ( C) Temperature ( C) A square millimetre per second (i.e., mm2/s) Parts per million (ppm) Insolubles Iron, silicon
Coded according to colour IP14 IP139 IP177 ASTM D93 ASTM D92 ASTM D92 IP71 ASTM D6304 PQ Analex Method ASTM D5185
with an HTF. Thus meaning that an HTF approved for incidental contact would not be needed or even relevant to their business. The second aim of this paper was therefore to assess the internal and external controls involved in the supply of food grade HTFs to determine if routine sampling and chemical analysis of HTFs had a place in the current framework.
was not defined because a number of HTF systems were sampled and chemically analysed by other companies prior to Global Heat Transfer's involvement. This makes it difficult to assess the exact age of some HTFs. Furthermore, a number of interventions are used so sustain the condition of an HTF, for example percentage dilution, which impacts the reliability and validity of assessing age.
2. Methods
2.1. Sampling of heat transfer fluids
The source of data in this paper came from retrospective test reports gathered by Global Heat Transfer Ltd. between 2008 and 2013. Data was analysed if: i. A report contained a non-food grade HTF (specifically Globaltherm™ M) or a food grade HTF (specifically Globaltherm™ FG) [5,14]. These HTFs were chosen to ensure the HTFs being analysed were consistent in terms of chemical composition and physical properties, as well as being sourced from the same supply. Table 3 summarises the physical properties of the food grade and non-food grade HTF analysed in this paper. ii. There was at least one test report between 2008 and 2013. iii. Indirect heating systems were sampled. All test results were analysed and analysis did not factor the make of the HTF system or age of the HTF or HTF system. The age
All identified systems were sampled whilst the system was operating under normal operating conditions. A 500 ml sample of the HTF was withdrawn from the HTF system using a custom designed sampling device that isolated the sampled HTF so a true and representative sample of the HTF was obtained. This procedure is described in detail elsewhere [7,8]. 2.2. Chemical analysis of heat transfer fluids Table 2 summarises the test parameters and test methods routinely conducted in the laboratory by Global Heat Transfer. These tests were conducted according to ISO14001 and ISO17025 standards [15,16]. The key parameters measured were carbon residue, total acid number (TAN), closed flash temperature, open flash temperature, kinematic viscosity, water content, ferrous wear debris (insolubles) and elements (e.g., iron and silicon).
Table 3 Characteristics of virgin food grade and non-food grade mineral heat transfer fluids. Parameter
Unit
Globaltherm™ M*
Globaltherm™ FG**
Base liquid
Description
#Mineral oil base
Grade
Description
Non-food grade
Synthetic base,a hydro treated, highly refined mineral oil base Food grade
Other products with similar characteristics
Description
BP Transcal N Castrol Perfecto HT Gulftherm 32 Shell HT S2
Calflo Purity FG Multitherm PG1 Marlotherm FG Paratherm NF Therminol XP
Appearance
Description
Colourless transparent liquid with no odour
Operating range Closed flash point Open flash point Kinematic viscosity at 40 C Auto-ignition point Pour point Maximum film temperature Boiling point at 1013 mbar
Viscous clear-yellow liquid with a mild odour 10 to 320 210 230 29.8 350 12 330 365
C C C mm2/s C C C C
20 to 326 210 216 29.8 350 29 343 371
* and **, values extracted from Globaltherm™ M and Globaltherm™ FG product datasheets [5,14]. Note: values for carbon residue, total acid number, strong acid number, fire point, water content, ferrous wear and insoluble were not reported in the product datasheets. Values for virgin HTFs are expected to be: carbon reside, 0% weight; total acid number, <0.05 mg KOH/g; strong acid number, 0 mg KOH/g; fire point, ~255 C; water content, <100 ppm; ferrous wear, <10; insolubles less than 5 microns, 0 ppm. a Extracted from Ref. [3].
C.I. Wright et al. / Applied Thermal Engineering 84 (2015) 94e103
2.3. Data analysis Data was analysed using a number of statistical tests. A Pearson's correlation test was used to assess the linearity between two parameters [8]. The Pearson correlation coefficient (r) returns a score with þ1 indicating a positive correlation; 1 indicating a negative correlation; and, 0 indicating no correlation. The linearity was tested for significance to determine if a true correlation existed (P < 0.05) or not (P 0.05). Correlation coefficient values are reported along with 95% confidence intervals (CI), which were calculated using a 2-sided Fisher's Z test, as well as degrees of freedom (DF). A chi-squared test (c2) was used to assess the frequency and distribution of ratings (i.e., satisfactory, caution, action and serious) of parameters and to determine if categorised parameters were related or not. In some cases test parameters values were missing. Data is presented as mean ± standard deviation (SD) unless otherwise stated. All analyses and calculations were conducted using Microsoft Excel 2007 and/or Analyse-it® version 3.71. Statistical significance was demonstrated when the P-value was taken at the 5% level (i.e., P < 0.05). The number of post-hoc comparisons was factored into the linear correlation test to account for the multiple comparisons [17].
97
found between carbon residue and TAN. Further comparisons revealed a number of moderate (r ±0.25 and <±0.5; P < 0.05) and weak correlations (r < ±0.25; P < 0.05) that were not consistent for both HTFs. These correlations are detailed below. 3.1.2.1. Non-food grade heat transfer fluid. Moderate positive linear correlations were found when kinematic viscosity was compared with carbon residue and TAN. Weak negative linear correlations were found when elements were compared with carbon residue and TAN. Weak negative correlations were found when TAN was compared with fire point and open flash point. 3.1.2.2. Food grade heat transfer fluid. Weak negative correlations were found when TAN was compared with open and closed flash points. 3.1.3. The relationships between fluid type, carbon residue and total acid number as determined by frequency distribution Table 4 highlights that mean and peak values for carbon residue and TAN were higher for the non-food grade HTF. Chi-squared tests were therefore used to assess: i) the association between ratings of carbon residue and TAN by HTF type; ii) the association between ratings of carbon residue by HTF type; and, iii) the association between ratings of TAN and HTF type (see Tables 6 and 7).
3. Results A total of 1223 HTF samples were included in this analysis e 943 mineral HTF and 280 food grade HTF. These were taken from 97 customers (24 food grade HTF and 73 mineral HTF) with multiple HTF system lines based at multiple locations in the UK and Ireland. 3.1. Comparison of non-food grade and food grade heat transfer fluids 3.1.1. Group data obtained for non-food grade and food grade heat transfer fluids Table 4 compares the mean ± SD and ranges for all measured parameters. The greatest differences between these HTFs was observed for carbon residue (0.12% weight or 118.0%); TAN (0.05 mg KOH/g or 40.0%); closed flash point (16.9 C or 12.1%); water content (þ49.0 ppm or þ35.6%); and, elements (10.3 ppm less than 5 microns or 186.9%). These parameters all differed by more than 10% with water content being the only parameter that was higher for the food grade HTF. 3.1.2. Linear comparison of test parameters Table 5 presents the correlations between parameters. For both HTFs, strong positive linear correlations (r ±0.5; P < 0.05) were
3.1.3.1. The frequency distribution of ratings of carbon residue and TAN by fluid type. In keeping with the linear correlations reported above, a significant association was found between the ratings of carbon residue and TAN (P < 0.05; Table 6a and b). Indeed, for both HTFs, the overall rating (i.e., totals for rows and columns) for carbon residue and TAN decreased as the overall condition of the HTF worsened. Table 6a and b was analysed to determine if individual cells in these contingency tables were greater than the critical value (i.e., 22.46). For the non-food grade HTF there were three cells that exceeded the critical value: 143.1 (carbon residue rating, 0.75 to <1.0; TAN rating, 0.4 to <1.0); 43.3 (carbon residue rating, 1.0; TAN rating, 0.4 to <1.0); and, 31.4 (carbon residue rating, 0.5 to <0.75; TAN rating, 0.2 to <0.4). For the food grade HTF only one cell exceeded the critical value: 107.2 (carbon residue rating, 1.0; TAN rating, 0.4 to <1.0). Thus suggesting an association between higher ratings of carbon residue (0.75 for non-food HTF; 1.0 for food grade HTF) and TAN (0.4 to <1.0 for both HTFs). The last point to make is the percentage of HTF systems that had a satisfactory rating. Table 6a and b revealed that 67% (631 of 936) of non-food grade HTF systems had a satisfactory rating (i.e., carbon residue <0.5; TAN <0.2) compared with 81% (224 of 278) of food
Table 4 Analysis results obtained for non-food grade and food grade heat transfer fluids (HTFs) e grouped data. Variable analysed
Non-food grade HTF (n ¼ 943)
Food grade HTF (n ¼ 279)
Absolute difference, % (food grade minus non-food grade)
Carbon residue (% weight) Total acid number (mg KOH/g) Strong acid number (mg KOH/g) Closed flash point ( C) Open flash point ( C) Fire point ( C) Kinematic viscosity at 40 C (mm2/s) Water content (ppm) Ferrous wear debris Elements (ppm < 5 microns)
0.23 ± 0.25 (range, 0.01e2.94) 0.18 ± 0.16 (range, 0.01e2.34) 0.00 ± 0.00 (range, 0.00e0.00) 156.3 ± 33.1 (range, 46.0e268.0) 203.1 ± 18.7 (range, 20.0e254.0) 241.52 ± 14.83 (range, 84.0e271.0) 30.5 ± 4.1 (range, 3.1e85.6)
0.10 ± 0.19 (range, 0.00e1.76) 0.13 ± 0.09 (range, 0.01e0.55) 0.00 ± 0.00 (range, 0.00e0.00) 139.4 ± 34.8 (range, 34.0e211.0) 199.4 ± 18.8 (range, 129.0e235.0) 244.0 ± 22.8 (range, 180.0e547.0) 31.4 ± 2.6 (range, 18.5e43.0)
0.12, 118.0 0.05, 40.0 0.00, 0.00 16.9, 12.1 3.7, 1.9 0.6, 0.3 0.9, 2.9
88.7 ± 266.6 (range, 10.0e4200.0) 14.4 ± 18.6 (range, 8.0e217.0) 15.9 ± 28.9 (range, 0.0e307.0)
137.7 ± 711.9 (range, 22.0e10,000) 15.9 ± 39.9 (range, 0.0e632.0) 5.5 ± 12.3 (range, 0.0e93.0)
49.0, 35.6 1.4, 9.0 10.3, 186.9
Means ± SD.
98
C.I. Wright et al. / Applied Thermal Engineering 84 (2015) 94e103
Table 5 Linear comparison of test parameters recorded for a non-food grade and food grade heat transfer fluid (HTF). Parameter a. Non-food grade HTF Carbon
TAN
b. Food grade HTF Carbon
TAN
Comparison
Pearson correlation coefficient (r)
Fisher 95% CI
DF
P-value
vs. vs. vs. vs. vs. vs. vs. vs.
TAN closed flash point open flash point elements ferrous wear debris kinematic viscosity water content fire point
þ0.680 0.012 0.068 þ0.190 þ0.060 þ0.360 þ0.010 0.067
0.640 to 0.709 0.076 to 0.052 0.131 to 0.004 0.132 to 0.255 0.003 to 0.124 0.300 to 0.412 0.050 to 0.079 0.132 to 0.001
934 934 933 934 934 934 921 887
<0.0001y 0.709 0.038 <0.0001y 0.063 <0.0001y 0.663 0.047
vs. vs. vs. vs. vs. vs. vs.
fire point water content kinematic viscosity ferrous wear debris elements open flash point closed flash point
0.091 þ0.030 þ0.270 þ0.050 þ0.140 0.091 0.011
0.156 to 0.025 0.039 to 0.090 0.206 to 0.325 0.010 to 0.117 0.081 to 0.206 0.154 to 0.027 0.075 to 0.053
889 924 937 938 937 935 936
0.007z 0.435 <0.0001z 0.099 <0.0001z 0.005z 0.747
vs. vs. vs. vs. vs. vs. vs. vs.
TAN closed flash point open flash point elements ferrous wear debris kinematic viscosity water content fire point
þ0.510 þ0.000 0.091 þ0.160 þ0.060 þ0.160 þ0.050 0.062
0.416 to 0.591 0.114 to 0.121 0.206 to 0.028 0.40 to 0.270 0.060 to 0.175 0.043 to 0.272 0.074 to 0.162 0.181 to 0.059
276 277 275 276 277 277 274 263
<0.0001y 0.950 0.133 0.009 0.332 0.007 0.459 0.319
vs. vs. vs. vs. vs. vs. vs.
fire point water content kinematic viscosity ferrous wear debris elements open flash point closed flash point
0.109 þ0.040 0.041 þ0.110 þ0.010 0.179 0.161
0.227 0.076 0.158 0.008 0.109 0.291 0.274
262 273 276 276 275 274 276
0.077 0.484 0.491 0.069 0.884 0.003z 0.007z
to to to to to to to
0.012 0.160 0.077 0.224 0.127 0.062 0.044
y (n ¼ 8) and z (n ¼ 7) comparisons were significant even when the number of post-hoc comparisons was factored into the analysis (i.e., P < 0.05 divided by the number of posthoc comparisons) [17]. TAN, total acid number. DF, degrees of freedom.
grade HTF systems. Further analysis revealed that 69% (648 of 936) and 81% (225 of 278) of non-food grade and food grade HTF systems had satisfactory ratings for TAN. These satisfactory ratings were even higher for carbon residue where satisfactory ratings were 80% (840 of 936) and 98% (272 of 278) for non-food grade and food grade HTFs respectively. 3.1.3.2. The relationship between carbon residue and fluid type. The chi-squared test showed a significant relation between carbon residue and HTF type (P < 0.05; Table 7a). Furthermore, the contingency table showed the frequency of satisfactory ratings was highest and a serious rating was the lowest. For the non-food grade HTF, the number of satisfactory ratings was lower than expected (i.e., observed minus expected) and higher than expected for caution, action and serious ratings. This was completely the opposite for the food grade HTF. 3.1.3.3. The relationship between total acid number and fluid type. This analysis revealed the same trends for non-food grade and food grade HTFs as was found for carbon residue (please see Table 7b). 3.2. The internal and external controls involved in the supply of food grade 3.2.1. Audits and controls involved in the supply of food grade HTFs A number of regulations e such as REACH [6] and CLP [10] e ensure that chemicals are distributed with sufficient warnings so they are safely managed and used (see Table 8). In the context of
food processing, HACCP [18] and quality/distribution audits [19] are used to ensure that food is managed safely both during processing and then when being distributed to the consumer. An additional safety measure is provided by trade bodies such as the British Retail Consortium [20]. This trade association is involved in all types of retail including food retailers (e.g., supermarkets) and promote responsible retailing. Furthermore, industrial insurers, such as FM Global [13], work closely with manufacturers to make sure that commercial operations are adequately insured. The manufacturer is central to all the activities defined in Fig. 2 yet there is no mechanism to ensure that food grade HTFs are used in the processing of food. Even though at this point in the process there is the potential for the food to come into contact with the HTF. The HACCP is an internal document used to define the acceptable contamination limits for a food, but the manufacturer may choose to dispose of all food that comes into contact with an HTF. For this reason a manufacturer may choose to use a non-food grade HTF as opposed to a food grade HTF. In another scenario, the retailer may stipulate the use of a food grade HTF and confirm its use through audits of the manufacturing facility. However, the HTF would not be sampled and chemically analysed and confirmation would be based on paperwork. Lastly, the insurer could stipulate the use of a food grade HTF, but, again, confirmation that a food grade HTF is used would be based on a paper-based audit. Insurers also assess the frequency that an HTF is sampled and chemically analysed to ensure the flash point temperature of the HTF is being managed as stipulated [7e9]. However, at no point would the insurer call request the HTF to be sampled and chemically analysed to validate documentation.
C.I. Wright et al. / Applied Thermal Engineering 84 (2015) 94e103
99
Table 6 The frequency distribution of carbon residue and total acid number (TAN) for non-food grade and food grade heat transfer fluids (HTFs). Calculation
TAN rating
Carbon residue rating
Totals
<0.5
0.5 to <0.75
0.75 to <1.0
1.0
<0.2
631.0 581.5 49.5 2446.4 4.2
12.0 30.5 18.5 340.8 11.2
2.0 20.8 18.8 352.3 17.0
3.0 15.2 12.2 149.6 9.8
648.0
Observed (O) Expected (E) OE O E2 (O E2)/E
0.2 to <0.4
174.0 193.8 19.8 393.9 2.0
28.0 10.2 17.8 318.5 31.4
8.0 6.9 1.1 1.2 0.2
6.0 5.1 0.9 0.9 0.2
216.0
Observed (O) Expected (E) OE O E2 (O E2)/E
0.4 to <1.0
35.0 61.9 26.9 724.9 11.7
4.0 3.2 0.8 0.6 0.2
20.0 2.2 17.8 316.4 143.1
10.0 1.6 8.4 70.2 43.3
69.0
Observed (O) Expected (E) OE O E2 (O E2)/E
1.0
0.0 2.7 2.7 7.2 0.0
0.0 0.1 0.1 0.0 0.0
0.0 0.1 0.1 0.0 0.0
3.0 0.1 2.9 8.6 0.0
3.0
840.0
44.0
30.0
22.0
936.0 274.2 9 22.46 <0.001
<0.2
224.0 220.1 3.9 14.9 0.1
1.0 2.4 1.4 2.0 0.8
0.0 0.8 0.8 0.7 0.8
0.0 1.6 1.6 2.6 1.6
225.0
Observed (O) Expected (E) OE O E2 (O E2)/E
0.2 to <0.4
45.0 47.0 2.0 3.9 0.1
2.0 0.5 1.5 2.2 4.2
1.0 0.2 0.8 0.7 4.0
0.0 0.3 0.3 0.1 0.3
48.0
Observed (O) Expected (E) OE O E2 (O E2)/E
0.4 to <1.0
3.0 4.9 1.9 3.6 0.7
0.0 0.1 0.1 0.0 0.1
0.0 0.0 0.0 0.0 0.0
2.0 0.0 2.0 3.9 107.2
5.0
Observed (O) Expected (E) OE O E2 (O E2)/E
1.0
0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0
0.0
272.0
3.0
1.0
2.0
278.0 120.0 9 22.46 <0.001
a. Non-food grade HTF Observed (O) Expected (E) OE O E2 (O E2)/E
Totals c2 value Degrees of freedom Critical value P-value b. Food grade HTF Observed (O) Expected (E) OE O E2 (O E2)/E
Totals c2 value Degrees of freedom Critical value P-value
Condition ratings for carbon residue: satisfactory, <0.5; caution, 0.5 to <0.75; action, 0.75 to <1.0; and, serious, 1.0. Condition ratings for TAN: satisfactory, <0.2; caution, 0.2 to <0.4; action, 0.4 to <1.0; and, serious, 1.0. Grey shaded area represents concurrent condition ratings for carbon residue and TAN.
4. Discussion Mean data clearly showed mean and maximum values for carbon residue and TAN were higher for non-food grade HTF than for the food grade HTF. Individual data was then explored and linear analysis showed there was a strong positive relationship between carbon residue and TAN for both HTFs. Subsequently, frequency distribution analysis, using chi-squared tests, suggested a possible association between high ratings of carbon residue (0.75 for non-
food HTF; 1.0 for food grade HTF) and relatively moderate ratings of TAN (0.4 to <1.0 for both HTFs). Meaning that carbon residue tended to be higher relative to TAN. The effect of HTF type was explored and showed that satisfactory ratings were highest and serious ratings were lowest and this was true for both HTFs. The number of satisfactory ratings was higher than expected for the food grade HTF and lower than expected for the non-food grade HTF. Thus meaning that non-food grade HTFs had a relatively higher number of scores in other categories judged to be not
100
C.I. Wright et al. / Applied Thermal Engineering 84 (2015) 94e103
Table 7 Determination of the relationship between fluid type and carbon residue and total acid number. HTF
Calculation
Carbon residue rating <0.5
0.75 to <1.0
1.0
45.0 37.0 8.0 64.1 1.7
30.0 23.9 6.1 37.3 1.6
23.0 20.0 3.0 8.8 0.4
938.0
272.0 254.9 17.1 291.4 1.1
3.0 11.0 8.0 64.1 5.8
1.0 7.1 6.1 37.3 5.2
3.0 6.0 3.0 8.8 1.5
279.0
1112.0
48.0
31.0
26.0
1217.0 15.8 3 12.84 <0.005
a. The relationship between carbon residue rating and fluid type Non-food grade Observed (O) 840.0 Expected (E) 857.1 OE 17.1 2 OE 291.4 (O E2)/E 0.3 Food grade
Observed (O) Expected (E) OE O E2 (O E2)/E
Totals c2 value Degrees of freedom Critical value P-value HTF
Calculation
TAN rating <0.2
Observed (O) Expected (E) OE O E2 (O E2)/E
Totals c2 value Degrees of freedom Critical value P-value
Totals 0.2 to <0.4
0.4 to <1.0
1.0
222.0 209.1 12.9 165.2 0.8
69.0 57.1 11.9 141.4 2.5
3.0 2.3 0.7 0.5 0.2
940.0
224.0 198.6 25.4 646.6 3.3
49.0 61.9 12.9 165.2 2.7
5.0 16.9 11.9 141.4 8.4
0.0 0.7 0.7 0.5 0.7
278.0
870.0
271.0
74.0
3.0
1218.0 19.4 3 16.27 <0.001
b. The relationship between total acid number (TAN) rating and fluid type Non-food grade Observed (O) 646.0 Expected (E) 671.4 OE 25.4 O E2 646.6 (O E2)/E 1.0 Food grade
Total 0.5 to <0.75
n, number of samples. NA, not applicable. Note: that in a small number of cases, values were missing and this explains why the totals here vary from those presented in the previous table. Condition ratings for carbon residue: satisfactory, <0.5; caution, 0.5 to <0.75; action, 0.75 to <1.0; and, serious, 1.0. Condition ratings for TAN: satisfactory, <0.2; caution, 0.2 to <0.4; action, 0.4 to<1.0; and, serious, 1.0.
Table 8 The intention and focus of various controls in the processing of food. Environment
Controls
Intention
Focus
Internal
SOP
Consistency of services and/or products
Internal
HACCP
External
REACH regulation
External
CLP legislation
External
Insurer
External
Quality and distribution audits
External
Consumer agencies
The documented processes that a company has in place and uses on a day-to-day basis to ensure services and/or products are delivered consistently across the organisation A system that helps food business operators look at how they handle food and introduces procedures to make sure the food produced is safe to eat Companies must identify and manage the risks linked to the substances they manufacture and market in the European Union. It applies to all chemical substances from those used in industrial processes to everyday fast-moving-consumer-goods. Companies need to demonstrate how to safely use the substances and this must be communicated to the users Ensures that the substances and mixtures placed on the market, across the globe, are classified, labelled and packaged appropriately Commercial and industrial insurers are used to manage risk needs (i.e., identification, assessment, avoidance and reduction) and protect the value created by the customer's organisation Examples include attainment and maintenance of ISO certification such as ‘ISO22000:2005 e Food Safety Management Systems’ which is a standard outlined for the control of food safety For example the BRC which is the UK trade association for the entire retail industry and includes food products sold to consumers
Managing food safely
Identification and management of risks associated with substances
Correct classification, labelling and packaging Managing risk
Managing food safely
Consumer safety
SOP, standard operating procedure; HACCP, hazard analysis and critical control point; REACH, registration, evaluation, authorisation and restriction of chemicals; CLP, classification, labelling and packaging; and, BRC, British retail consortium.
C.I. Wright et al. / Applied Thermal Engineering 84 (2015) 94e103
101
satisfactory. Finally, satisfactory ratings for carbon residue and TAN were markedly higher for food grade HTFs: carbon residue, 90% versus 98% (non-food versus food grade HTFs respectively); and, TAN, 69% versus 81%. 4.1. Comparison of non-food grade and food grade heat transfer fluids It is estimated that there are around 4000 companies using high temperature HTF systems in the UK and Ireland [21] with roughly one-in-five systems, based on the HTF systems that the authors is aware of, using HTFs to process food. Food grade HTFs are used in food processing as they are generally regarded as being safe for incidental contact. Another feature of food grade HTFs is that they are non-fouling and the analysis of this characteristic was the primary goal of this paper. Non-fouling is used to describe the carbon formatting capacity of an HTF. If carbon residue levels are reduced this means that there is less carbon present which means that less carbon adheres to the internal surfaces of an HTF system. The authors were unaware of comparisons between non-food grade and food grade HTFs so these HTFs were compared to assess mean carbon residue levels and ranges, and how these were distributed (i.e., if they are satisfactory or not). The first aspect of this analysis showed that carbon level was 0.10 and 0.23 percentage weight for food grade and non-food grade HTF. This reflects a percentage difference of 118% (0.12/0.10 * 100). TAN levels were also lower for food grade HTF (0.13 versus 0.18; food grade versus non-food grade HTF). Other differences between HTFs were identified (see Table 4). These included closed flash point, water content and elements, the latter two reflecting contamination and not a physical difference in the HTFs. The lower carbon residue level for food grade HTFs is very interesting because during thermal degradation of an HTF two products are formed e long (commonly referred to as ‘heavy ends’) and short chain (‘light ends’) hydrocarbons [9]. Testing is used to identify these components and usually seen as increased carbon residue and decreased closed flash point, which decreases as the composition of light ends in the HTF increases. With this in mind, it is interesting that non-food grade HTF had a higher carbon residue level but this was accompanied by a currently lower closed flash point as might have been expected. The second aspect of the analysis was to explore the relationship between the test parameters (see Table 5). This was done by constructing xey plots and assessing the correlation between parameters. The principle finding was the strong positive (r ±0.5; P < 0.05) correlation between carbon residue and TAN. This was seen for both HTFs. Other correlations revealed weaker correlations (r < ±0.5; P < 0.05). Interestingly, closed and open flash points were found to negatively relate to TAN for food grade HTF. For non-food grade HTF TAN was found to negatively relate to fire point and open flash point, but not closed flash point. This data demonstrates the existence of subtle differences between HTFs but also shows that the one consistent finding between HTFs was the existence of a relationship between carbon residue and TAN. This therefore formed the basis of the current comparison between non-food and food grade HTFs. Fig. 1 compares the plots for non-food grade and food grade HTFs. The equation for inserted lines are y ¼ 0.4435x þ 0.075 for non-food grade and y ¼ 0.2809x þ 0.0984 for the food grade HTF. So the question here is what is the relevance of this data? The equations suggest that for any given level of carbon residue, the TAN is lower for food grade HTFs. Conversely, this also means that for any given level of TAN, the carbon residue level is higher for non-food grade HTF. To understand how test report scores were distributed, carbon residue and TAN were categorised into four groups. This categorisation had a maximal score of 1 and reflected a serious level for
Fig. 1. Linear comparisons of carbon residue and total acid number (TAN) for a) nonfood grade and b) food grade heat transfer fluids (HTFs).
both parameters. Whereas a score close to zero reflected that seen for a virgin HTF. These categorisations are reported in Tables 6 and 7. This specific analysis was done to understand if carbon residue and TAN varied by HTF type and if this rating explained the lower mean carbon residue and TAN scores for food grade HTF. Initially HTFs were analysed separately so that carbon residue and TAN
Fig. 2. An illustration of the audits (internal and external) and controls involved in the food processing supply chain. SOP, standard operating procedure; HACCP, hazard analysis and critical control point; REACH, registration, evaluation, authorisation and restriction of chemicals; CLP, classification, labelling and packaging; and, BRC, British retail consortium.
102
C.I. Wright et al. / Applied Thermal Engineering 84 (2015) 94e103
could be compared by HTF type, and then together to see if carbon residue and TAN differed by HTF type. This analysis revealed an association between irrespective of how the data was analysed. Indeed, both approaches used chi-squared tests and showed that for both HTFs the number of satisfactory ratings (i.e., <0.5 for carbon residue and <0.2 for TAN) was higher than expected. Furthermore, the number of unsatisfactory ratings (i.e., any category that was not judged to be satisfactory) decreased as the condition rating progressive worsened. Thus meaning that the smallest tally of test results existed in the category where carbon residue and TAN was 1.0. Analysis also suggests a possible association between high ratings of carbon residue (0.75 for non-food HTF; 1.0 for food grade HTF) and relatively moderate ratings of TAN (0.4 to <1.0 for both HTFs). Meaning that carbon residue tended to be higher relative to TAN, which is also seen in the linear correlation analysis (see Fig. 1). Chi-squared testing was also used to assess the association between carbon residue and both HTFs (see Table 7). This showed that the number of test reports that had a satisfactory rating was higher than expected in the food grade HTF group and lower than expected in the non-food grade HTF group. Furthermore, the tallies of test reports with an unsatisfactory rating were higher for the non-food grade HTF and lower for the food grade HTF. The same finding was also true for assessments of TAN. This clearly shows that food grade HTF had more satisfactory ratings and fewer serious ratings than the non-food grade HTF. 4.2. The internal and external controls involved in the supply of food grade The second aim was to assess the landscape for audits and controls encompassed in the supply of food grade HTFs. The supply of food grade HTFs is highly regulated in Europe and companies must comply with REACH and CLP regulations [6,10]. These regulations also concern the safe distribution and handling of HTFs [6,10]. Other elements include the processing controls and safeguards such as HACCP [18] and quality/distribution audits [19]. The British Retail Consortium is a trade association [20] that is involved in the protection of the consumer and promotion of responsible retailing. Lastly, industrial insurers [13] that work to ensure HTFs and HTF systems are managed safely. In addition, a manufacturer should only source food grade HTFs that are suitable for food processing and incidental contact with food [17]. HT-1 certification is governed by two well-known bodies e the NSF and InS. In the case of the NSF, the components comprising a fluid are compared against the NSF's registration guidelines for proprietary substances and non-food compounds and 21 CFR [18]. Based on a toxicologists' analysis, an HTF is then registered as HT-1 [11], which is specific to food grade HTFs. In some cases, but not all, the use of food grade HTFs is stipulated by insurers and food retailers, and certain manufacturers will be routinely audited to ensure that an appropriate HTF is being used in the processing of food. However, the current framework means there is potential for companies to use HTFs not suitable for food processing. One justification being that a company may choose to dispose all foods if they come into contact with the HTF they are using. Thus meaning that an HTF approved for incidental contact would not be needed or even relevant to their business. Another example for not using food grade HTFs is that the HTF will not come into contact with food, which seems logical but caution should be taken as leaks can happen and go undetected until the consumer comes into contact with the processed food. Indeed, in 1998 more than 222,260 kg of smoked boneless hams were recalled by Smithfield Foods because they were contaminated by a gear lubricant after several customers reported a ‘bad taste’ and ‘burning in their throat’ which lasted anywhere up to 3 h [22]. The key take
home message being that an inappropriate HTF was used and that these adverse events could have been minimised if a food grade HTF had been used. A review of all processes, audits and controls indicates that there is a gap in current practice and this means that non-food grade HTFs may be used or that the HTF is never sampled and chemically analysed to confirm it is a food grade HTF. Such testing is quick and easy to conduct [7] and can be done without interrupting production. Indeed, it is possible for HTF systems to be sampled prior to and following any intervention that affects the HTF [8] and whilst the HTF system is running under normal conditions. Such testing should be followed-up with routine and regular HTF sampling to assess the effectiveness of any intervention and also to assess the ongoing state of the HTF [23]. A number of interventions can be used to help maintain the condition of the HTF such as the installation of a light-ends removal kit (LERK) which can be used be fitted permanently to the HTF system or used intermittently [9] in a similar manner to batch venting but without the need to interrupt system production [24]. Testing can also be used to assess the effectiveness of such interventions and to propose subsequent actions. It would seem logical that HTF sampling is performed by a specialist in this area to deter malpractice. The test results from such testing could then be provided to: 1) the insurer to show the HTF is being managed and the HTF system is safe; 2) retailers to demonstrate that an appropriate food grade HTF is being used in food processing; and, 3) external agencies to show that consumer safety is not being compromised. 5. Conclusions The primary aim of this paper was to compare the non-fouling nature of non-food grade and food grade HTFs. The take home message from this analysis is that non-fouling is observed as lower carbon residue and TAN levels. Indeed, mean carbon residue and TAN levels were higher for non-food grade HTF, but irrespective of this, carbon residue and TAN were found to be strongly related for both HTFs. Linear plots demonstrate that carbon residue was higher relative to TAN. Analysis looking at the frequency of distribution shows that ratings for HTFs were highest (i.e., observed tallies were higher than expected) in the category judged to be satisfactory. The same analysis technique also showed that food grade HTFs tended to have more scores in the satisfactory category rating than nonfood grade HTFs. Conversely, non-food grade HTF had more unsatisfactory ratings whereas the tallies for food grade HTF were the complete opposite (i.e., lower than expected). In fact, comparisons showed that in 69% and 81% of cases TAN was <0.2 (for non-food grade and food grade HTF, respectively). This difference is þ12% in favour of the food grade HTF. In 90% and 98% of cases, carbon residue was <0.5 e a difference of þ8% in favour of food grade HTF. The secondary aim was to analyse the legal, environmental and regulatory landscape concerning the use of food grade HTF in food processing. This revealed that there is no body physically checking that a food grade HTF is being used in the HTF system. One easy solution to consider is the random and independent sampling and chemical analysis of HTFs to ensure that a food grade HTF is being used and this would also work as an additional measure underpinning consumer safety. Acknowledgements The author would like to acknowledge the Global Heat Transfer team. Specifically the engineering team (Danny Bradford, Dave Dyer and Martyn Tinsley) and the technical support from Andy Burns, Lisa Cho and Ian Halliwell.
C.I. Wright et al. / Applied Thermal Engineering 84 (2015) 94e103
References [1] ReportsnReports.com Updates its Store with “Global Heat Transfer Fluid Market [Thermic Fluid/High Temperature/Synthetic Heat Transfer Fluids] By Product Type [Mineral Oils, Silicone & Aromatics, PAG & Glycol Based Products & Others], Applications & Geography e Forecasts To 2017” Market Research Report, Source: www.marketsandmarkets.com (accessed 22.10.14). [2] D.J. Kukulka, M. Devgun, Fluid temperature and velocity effect on fouling, Appl. Therm. Eng. 27 (2007) 2732e2744. [3] Wagner O. Walter, Heat transfer technique with organic media, in: Heat Transfer Media, second ed., Maria-Eich-Strabe, Graefelfing, Germany, 1997, pp. 4e58 (Chapter 2). [4] White Mineral Oil (petroleum), Source: http://apps.echa.europa.eu/registered/ data/dossiers/DISS-9ea08dfc-55a5-3699-e044-00144f67d031/DISS-9ea08dfc55a5-3699-e044-00144f67d031_DISS-9ea08dfc-55a5-3699-e04400144f67d031.html (accessed 22.10.14). [5] Globaltherm™ FG, Source: Globalheattransfer.co.uk (accessed 30.10.14). [6] REACH regulation, Source: http://echa.europa.eu/regulations/reach/ understanding-reach (accessed 21.10.14). [7] C.I. Wright, A. Burns, C.J. Jones, Sampling hot heat transfer fluids: simple insights for gaining a representative sample, Int. J. Eng. Innov. Technol. 3 (2013) 202e204. [8] C.I. Wright, Thermal heat transfer fluid problems following a system flush with caustic and water, Case Stud. Therm. Eng. 2 (2014) 91e94. [9] C.I. Wright, Effective management of heat transfer fluid flash point temperatures using a light-ends removal kit (LERK), Case Stud. Therm. Eng. 4 (2014) 9e14. [10] CLP legislation, http://echa.europa.eu/regulations/clp/legislation (accessed 21.10.14). [11] P. Bird, Advice on how to convert your processing plant to H1 and HT-1 registered food machinery lubricants, Food Market. Technol. (June 2009). Source: http://d-nb.info/994769512/34 (accessed 22.10.14).
103
[12] CFR e Code of Federal Regulations Title 21, Source: www.accessdata.fda.gov/ scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr¼178.3570 (accessed 22.10.14). [13] FM Global, Source: www.fmglobal.com/page.aspx?id¼01010000 (accessed 21.10.14). [14] Globaltherm™ M, Source: Globalheattransfer.co.uk (accessed 30.10.14). [15] ISO 14001, Environmental Management, 2004. Source: http://www.iso.org/ iso/iso14000 (accessed 22.10.14). [16] ISO/IEC 17025, General Requirements for the Competence of Testing and Calibration Laboratories, 2005. Source: http://www.standards.org/standards/ listing/iso_17025 (accessed 22.10.14). [17] C.I. Wright, H.J. Scholten, J.C. Schilder, B.M. Elsen, W. Hanselaar, C.I. Kroner, R. Draijer, J.J. Kastelein, W. Stok, J. Karemaker, E. de Groot, Arterial stiffness, endothelial function and microcirculatory reactivity in healthy young males, Clin. Physiol. Funct. Imaging 28 (5) (2008) 299e306. [18] Hazard Analysis and Critical Control Point (HACCP), www.food.gov.uk/ business-industry/caterers/haccp (accessed 21.10.14). [19] ISO 22000 e Food safety management, Source: http://www.iso.org/iso/home/ standards/management-standards/iso22000.htm (accessed 21.10.14). [20] British Retail Consortium (BRC), Source: www.brc.org.uk/brc_home.asp (accessed 21.10.14). [21] A. McKay, R. Franklin, Good practice: fire and explosion hazards with thermal fluid systems, Loss Prev. Bull. 219 (2011) 13e17. [22] S. Gebarin, The basics of food-grade lubricants, Machin. Lubric. (2009). Source: www.machinerylubrication.com/Read/1857/food-grade-lubricants-basics (accessed 22.10.14). [23] C.I. Wright, E. Picot, T. Bembridge, The relationship between the condition of a mineral-based heat transfer fluid and the frequency that it is sampled and chemically analysed, Appl. Therm. Eng. 75 (2015) 918e922. [24] C.I. Wright, J. Premel, Heat transfer system safety: comparing the effectiveness of batch venting and a light-ends removal kit (LERK), Case Stud. Therm. Eng. 4 (2014) 215e221.